Spaces:
Running
Running
martin
commited on
Commit
·
67c46fd
1
Parent(s):
703c9a0
initial
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +7 -35
- .gitignore +1 -0
- Dockerfile +46 -0
- __init__.py +0 -0
- app.py +173 -0
- assets/assistant.png +3 -0
- assets/user.png +3 -0
- clone_hf_model.sh +57 -0
- cosyvoice/__init__.py +0 -0
- cosyvoice/cli/__init__.py +0 -0
- cosyvoice/cli/cosyvoice.py +68 -0
- cosyvoice/cli/frontend.py +106 -0
- cosyvoice/cli/model.py +32 -0
- cosyvoice/flow/decoder.py +238 -0
- cosyvoice/flow/flow.py +196 -0
- cosyvoice/flow/flow_matching.py +315 -0
- cosyvoice/flow/length_regulator.py +65 -0
- cosyvoice/hifigan/f0_predictor.py +55 -0
- cosyvoice/hifigan/generator.py +566 -0
- cosyvoice/matcha/audio.py +90 -0
- cosyvoice/matcha/decoder.py +511 -0
- cosyvoice/matcha/flow_matching.py +141 -0
- cosyvoice/matcha/transformer.py +443 -0
- cosyvoice/transformer/__init__.py +0 -0
- cosyvoice/transformer/activation.py +87 -0
- cosyvoice/transformer/attention.py +322 -0
- cosyvoice/transformer/convolution.py +147 -0
- cosyvoice/transformer/decoder.py +418 -0
- cosyvoice/transformer/decoder_layer.py +132 -0
- cosyvoice/transformer/embedding.py +293 -0
- cosyvoice/transformer/encoder.py +633 -0
- cosyvoice/transformer/encoder_layer.py +237 -0
- cosyvoice/transformer/label_smoothing_loss.py +98 -0
- cosyvoice/transformer/positionwise_feed_forward.py +116 -0
- cosyvoice/transformer/subsampling.py +391 -0
- cosyvoice/utils/__init__.py +0 -0
- cosyvoice/utils/audio.py +90 -0
- cosyvoice/utils/class_utils.py +78 -0
- cosyvoice/utils/common.py +169 -0
- cosyvoice/utils/executor.py +151 -0
- cosyvoice/utils/file_utils.py +49 -0
- cosyvoice/utils/frontend_utils.py +142 -0
- cosyvoice/utils/mask.py +226 -0
- cosyvoice/utils/scheduler.py +761 -0
- cosyvoice/utils/train_utils.py +350 -0
- funasr_detach/__init__.py +38 -0
- funasr_detach/auto/__init__.py +0 -0
- funasr_detach/auto/auto_frontend.py +90 -0
- funasr_detach/auto/auto_model.py +573 -0
- funasr_detach/auto/auto_tokenizer.py +7 -0
.gitattributes
CHANGED
|
@@ -1,35 +1,7 @@
|
|
| 1 |
-
*.
|
| 2 |
-
*.
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
-
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
-
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
-
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
-
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
-
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
-
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
-
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
-
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
-
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
-
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
-
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
-
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
-
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
-
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
-
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
-
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
-
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
-
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
-
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
-
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 29 |
-
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 30 |
-
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 31 |
-
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 32 |
-
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 33 |
-
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
-
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
-
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 1 |
+
*.png filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.wav filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
assets/user.png filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
assets/assistant.png filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
speakers/闫雨婷_prompt.wav filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
speakers/闫雨婷RAP_prompt.wav filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
speakers/闫雨婷VOCAL_prompt.wav filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
.gitignore
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
__pycache__/
|
Dockerfile
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
FROM nvidia/cuda:12.1.0-base-ubuntu20.04
|
| 2 |
+
|
| 3 |
+
ENV TZ=Asia/Shanghai
|
| 4 |
+
RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime \
|
| 5 |
+
&& echo $TZ > /etc/timezone
|
| 6 |
+
|
| 7 |
+
RUN apt-get update \
|
| 8 |
+
&& apt-get install -y build-essential \
|
| 9 |
+
&& apt-get install -y wget \
|
| 10 |
+
&& apt-get install -y software-properties-common curl zip unzip git-lfs awscli libssl-dev openssh-server vim \
|
| 11 |
+
&& apt-get install -y net-tools iputils-ping iproute2
|
| 12 |
+
|
| 13 |
+
RUN apt-get install --reinstall ca-certificates && update-ca-certificates
|
| 14 |
+
|
| 15 |
+
RUN add-apt-repository -y 'ppa:deadsnakes/ppa' && apt update
|
| 16 |
+
RUN apt install python3.10 python3.10-dev python3.10-distutils python3.10-venv -y \
|
| 17 |
+
&& apt-get clean \
|
| 18 |
+
&& rm -rf /var/lib/apt/lists/*
|
| 19 |
+
|
| 20 |
+
RUN wget -qO- https://bootstrap.pypa.io/get-pip.py | python3.10
|
| 21 |
+
RUN ln -s /usr/bin/python3.10 /usr/bin/python
|
| 22 |
+
RUN pip uninstall -y Pillow && pip install pillow
|
| 23 |
+
|
| 24 |
+
# https://huggingface.co/docs/hub/spaces-sdks-docker#permissions
|
| 25 |
+
RUN useradd -m -u 1000 user
|
| 26 |
+
USER user
|
| 27 |
+
|
| 28 |
+
ENV HOME="/home/user" \
|
| 29 |
+
PATH="/home/user/.local/bin:${PATH}"
|
| 30 |
+
|
| 31 |
+
RUN python3.10 -m pip install pipx
|
| 32 |
+
RUN pipx install poetry
|
| 33 |
+
|
| 34 |
+
RUN poetry --version || { echo 'Poetry installation check failed' ; exit 1; }
|
| 35 |
+
|
| 36 |
+
WORKDIR /workspace
|
| 37 |
+
|
| 38 |
+
COPY --chown=user requirements.txt .
|
| 39 |
+
RUN pip install -r requirements.txt
|
| 40 |
+
|
| 41 |
+
COPY --chown=user . .
|
| 42 |
+
|
| 43 |
+
RUN pip install gradio
|
| 44 |
+
RUN chmod +x clone_hf_model.sh
|
| 45 |
+
ENV HF_MODEL_PATH="/tmp/hf_model"
|
| 46 |
+
CMD ["./clone_hf_model.sh", "$HF_MODEL_PATH", "&&", "python", "app.py", "--model", "$HF_MODEL_PATH"]
|
__init__.py
ADDED
|
File without changes
|
app.py
ADDED
|
@@ -0,0 +1,173 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import time
|
| 3 |
+
from pathlib import Path
|
| 4 |
+
import torchaudio
|
| 5 |
+
from stepaudio import StepAudio
|
| 6 |
+
|
| 7 |
+
from funasr import AutoModel
|
| 8 |
+
from funasr.utils.postprocess_utils import rich_transcription_postprocess
|
| 9 |
+
|
| 10 |
+
CACHE_DIR = "/tmp/gradio/"
|
| 11 |
+
system_promtp = {"role": "system", "content": "适配用户的语言,用简短口语化的文字回答"}
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
class CustomAsr:
|
| 15 |
+
def __init__(self, model_name="iic/SenseVoiceSmall", device="cuda"):
|
| 16 |
+
self.model = AutoModel(
|
| 17 |
+
model=model_name,
|
| 18 |
+
vad_model="fsmn-vad",
|
| 19 |
+
vad_kwargs={"max_single_segment_time": 30000},
|
| 20 |
+
device=device,
|
| 21 |
+
)
|
| 22 |
+
|
| 23 |
+
def run(self, audio_path):
|
| 24 |
+
res = self.model.generate(
|
| 25 |
+
input=audio_path,
|
| 26 |
+
cache={},
|
| 27 |
+
language="auto", # "zh", "en", "yue", "ja", "ko", "nospeech"
|
| 28 |
+
use_itn=True,
|
| 29 |
+
batch_size_s=60,
|
| 30 |
+
merge_vad=True, #
|
| 31 |
+
merge_length_s=15,
|
| 32 |
+
)
|
| 33 |
+
text = rich_transcription_postprocess(res[0]["text"])
|
| 34 |
+
return text
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
def add_message(chatbot, history, mic, text, asr_model):
|
| 38 |
+
if not mic and not text:
|
| 39 |
+
return chatbot, history, "Input is empty"
|
| 40 |
+
|
| 41 |
+
if text:
|
| 42 |
+
chatbot.append({"role": "user", "content": text})
|
| 43 |
+
history.append({"role": "user", "content": text})
|
| 44 |
+
elif mic and Path(mic).exists():
|
| 45 |
+
chatbot.append({"role": "user", "content": {"path": mic}})
|
| 46 |
+
# 使用用户语音的 asr 结果为了加速推理
|
| 47 |
+
text = asr_model.run(mic)
|
| 48 |
+
chatbot.append({"role": "user", "content": text})
|
| 49 |
+
history.append({"role": "user", "content": text})
|
| 50 |
+
|
| 51 |
+
print(f"{history=}")
|
| 52 |
+
return chatbot, history, None
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
def reset_state():
|
| 56 |
+
"""Reset the chat history."""
|
| 57 |
+
return [], [system_promtp]
|
| 58 |
+
|
| 59 |
+
|
| 60 |
+
def save_tmp_audio(audio, sr):
|
| 61 |
+
import tempfile
|
| 62 |
+
|
| 63 |
+
with tempfile.NamedTemporaryFile(
|
| 64 |
+
dir=CACHE_DIR, delete=False, suffix=".wav"
|
| 65 |
+
) as temp_audio:
|
| 66 |
+
temp_audio_path = temp_audio.name
|
| 67 |
+
torchaudio.save(temp_audio_path, audio, sr)
|
| 68 |
+
|
| 69 |
+
return temp_audio.name
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
def predict(chatbot, history, audio_model):
|
| 73 |
+
"""Generate a response from the model."""
|
| 74 |
+
try:
|
| 75 |
+
text, audio, sr = audio_model(history, "闫雨婷")
|
| 76 |
+
print(f"predict {text=}")
|
| 77 |
+
audio_path = save_tmp_audio(audio, sr)
|
| 78 |
+
chatbot.append({"role": "assistant", "content": {"path": audio_path}})
|
| 79 |
+
chatbot.append({"role": "assistant", "content": text})
|
| 80 |
+
history.append({"role": "assistant", "content": text})
|
| 81 |
+
except Exception as e:
|
| 82 |
+
print(e)
|
| 83 |
+
gr.Warning(f"Some error happend, retry submit")
|
| 84 |
+
return chatbot, history
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
def _launch_demo(args, audio_model, asr_model):
|
| 88 |
+
with gr.Blocks(delete_cache=(86400, 86400)) as demo:
|
| 89 |
+
gr.Markdown("""<center><font size=8>Step Audio Chat</center>""")
|
| 90 |
+
chatbot = gr.Chatbot(
|
| 91 |
+
elem_id="chatbot",
|
| 92 |
+
avatar_images=["assets/user.png", "assets/assistant.png"],
|
| 93 |
+
min_height=800,
|
| 94 |
+
type="messages",
|
| 95 |
+
)
|
| 96 |
+
# 保存 chat 历史,不需要每次再重新拼格式
|
| 97 |
+
history = gr.State([system_promtp])
|
| 98 |
+
mic = gr.Audio(type="filepath")
|
| 99 |
+
text = gr.Textbox(placeholder="Enter message ...")
|
| 100 |
+
|
| 101 |
+
with gr.Row():
|
| 102 |
+
clean_btn = gr.Button("🧹 Clear History (清除历史)")
|
| 103 |
+
regen_btn = gr.Button("🤔️ Regenerate (重试)")
|
| 104 |
+
submit_btn = gr.Button("🚀 Submit")
|
| 105 |
+
|
| 106 |
+
def on_submit(chatbot, history, mic, text):
|
| 107 |
+
chatbot, history, error = add_message(
|
| 108 |
+
chatbot, history, mic, text, asr_model
|
| 109 |
+
)
|
| 110 |
+
if error:
|
| 111 |
+
gr.Warning(error) # 显示警告消息
|
| 112 |
+
return chatbot, history, None, None
|
| 113 |
+
else:
|
| 114 |
+
chatbot, history = predict(chatbot, history, audio_model)
|
| 115 |
+
return chatbot, history, None, None
|
| 116 |
+
|
| 117 |
+
submit_btn.click(
|
| 118 |
+
fn=on_submit,
|
| 119 |
+
inputs=[chatbot, history, mic, text],
|
| 120 |
+
outputs=[chatbot, history, mic, text],
|
| 121 |
+
concurrency_limit=4,
|
| 122 |
+
concurrency_id="gpu_queue",
|
| 123 |
+
)
|
| 124 |
+
clean_btn.click(
|
| 125 |
+
reset_state,
|
| 126 |
+
outputs=[chatbot, history],
|
| 127 |
+
show_progress=True,
|
| 128 |
+
)
|
| 129 |
+
|
| 130 |
+
def regenerate(chatbot, history):
|
| 131 |
+
while chatbot and chatbot[-1]["role"] == "assistant":
|
| 132 |
+
chatbot.pop()
|
| 133 |
+
while history and history[-1]["role"] == "assistant":
|
| 134 |
+
print(f"discard {history[-1]}")
|
| 135 |
+
history.pop()
|
| 136 |
+
return predict(chatbot, history, audio_model)
|
| 137 |
+
|
| 138 |
+
regen_btn.click(
|
| 139 |
+
regenerate,
|
| 140 |
+
[chatbot, history],
|
| 141 |
+
[chatbot, history],
|
| 142 |
+
show_progress=True,
|
| 143 |
+
concurrency_id="gpu_queue",
|
| 144 |
+
)
|
| 145 |
+
|
| 146 |
+
demo.queue().launch(
|
| 147 |
+
share=False,
|
| 148 |
+
server_port=args.server_port,
|
| 149 |
+
server_name=args.server_name,
|
| 150 |
+
)
|
| 151 |
+
|
| 152 |
+
|
| 153 |
+
if __name__ == "__main__":
|
| 154 |
+
from argparse import ArgumentParser
|
| 155 |
+
import os
|
| 156 |
+
|
| 157 |
+
parser = ArgumentParser()
|
| 158 |
+
parser.add_argument("--model-path", type=str, required=True, help="Model path.")
|
| 159 |
+
parser.add_argument(
|
| 160 |
+
"--server-port", type=int, default=7860, help="Demo server port."
|
| 161 |
+
)
|
| 162 |
+
parser.add_argument(
|
| 163 |
+
"--server-name", type=str, default="0.0.0.0", help="Demo server name."
|
| 164 |
+
)
|
| 165 |
+
args = parser.parse_args()
|
| 166 |
+
|
| 167 |
+
audio_model = StepAudio(
|
| 168 |
+
tokenizer_path=os.path.join(args.model_path, "Step-Audio-Tokenizer"),
|
| 169 |
+
tts_path=os.path.join(args.model_path, "Step-Audio-TTS-3B"),
|
| 170 |
+
llm_path=os.path.join(args.model_path, "Step-Audio-Chat"),
|
| 171 |
+
)
|
| 172 |
+
asr_model = CustomAsr()
|
| 173 |
+
_launch_demo(args, audio_model, asr_model)
|
assets/assistant.png
ADDED
|
Git LFS Details
|
assets/user.png
ADDED
|
Git LFS Details
|
clone_hf_model.sh
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/bin/bash
|
| 2 |
+
|
| 3 |
+
if [ -z "$HF_USER_NAME" ]; then
|
| 4 |
+
echo "错误:环境变量 HF_USER_NAME 未设置!"
|
| 5 |
+
exit 1
|
| 6 |
+
fi
|
| 7 |
+
|
| 8 |
+
if [ -z "$HF_USER_TOKEN" ]; then
|
| 9 |
+
echo "错误:环境变量 HF_USER_TOKEN 未设置!"
|
| 10 |
+
exit 1
|
| 11 |
+
fi
|
| 12 |
+
|
| 13 |
+
# 启用Git LFS支持
|
| 14 |
+
git lfs install --force
|
| 15 |
+
|
| 16 |
+
# 定义需要克隆的仓库列表
|
| 17 |
+
BASE_REPO_URL="https://${HF_USER_NAME}:${HF_USER_TOKEN}@huggingface.co/stepfun-ai"
|
| 18 |
+
REPOSITORIES=(
|
| 19 |
+
"Step-Audio-Tokenizer"
|
| 20 |
+
"Step-Audio-TTS-3B"
|
| 21 |
+
"Step-Audio-Chat"
|
| 22 |
+
)
|
| 23 |
+
|
| 24 |
+
# 定义本地存放仓库的目录,默认为当前目录
|
| 25 |
+
LOCAL_DIR="${1:-$(pwd)}"
|
| 26 |
+
|
| 27 |
+
# 克隆函数(带无限重试机制)
|
| 28 |
+
clone_with_retry() {
|
| 29 |
+
local repo_name=$1
|
| 30 |
+
local repo_url="${BASE_REPO_URL}/${repo_name}"
|
| 31 |
+
local target_dir="${LOCAL_DIR}/${repo_name}"
|
| 32 |
+
|
| 33 |
+
# 检查是否已存在目录
|
| 34 |
+
if [ -d "${target_dir}" ]; then
|
| 35 |
+
echo "目录 ${target_dir} 已存在,跳过克隆。"
|
| 36 |
+
return 0
|
| 37 |
+
fi
|
| 38 |
+
|
| 39 |
+
# 无限重试循环
|
| 40 |
+
while true; do
|
| 41 |
+
echo "正在尝试克隆 ${repo_name} 到 ${target_dir}..."
|
| 42 |
+
if git clone "${repo_url}" "${target_dir}"; then
|
| 43 |
+
echo "成功克隆 ${repo_name} 到 ${target_dir}"
|
| 44 |
+
return 0
|
| 45 |
+
else
|
| 46 |
+
echo "克隆失败, 5秒后重试..."
|
| 47 |
+
sleep 5
|
| 48 |
+
fi
|
| 49 |
+
done
|
| 50 |
+
}
|
| 51 |
+
|
| 52 |
+
# 遍历所有仓库进行克隆
|
| 53 |
+
for repo in "${REPOSITORIES[@]}"; do
|
| 54 |
+
clone_with_retry "${repo}"
|
| 55 |
+
done
|
| 56 |
+
|
| 57 |
+
echo "所有仓库已成功下载!"
|
cosyvoice/__init__.py
ADDED
|
File without changes
|
cosyvoice/cli/__init__.py
ADDED
|
File without changes
|
cosyvoice/cli/cosyvoice.py
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import os
|
| 15 |
+
import uuid
|
| 16 |
+
import time
|
| 17 |
+
from tqdm import tqdm
|
| 18 |
+
import torch
|
| 19 |
+
import torchaudio
|
| 20 |
+
from hyperpyyaml import load_hyperpyyaml
|
| 21 |
+
from cosyvoice.cli.frontend import CosyVoiceFrontEnd
|
| 22 |
+
from cosyvoice.cli.model import CosyVoiceModel
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
class CosyVoice:
|
| 26 |
+
|
| 27 |
+
def __init__(
|
| 28 |
+
self,
|
| 29 |
+
model_dir,
|
| 30 |
+
):
|
| 31 |
+
self.model_dir = model_dir
|
| 32 |
+
with open("{}/cosyvoice.yaml".format(model_dir), "r") as f:
|
| 33 |
+
configs = load_hyperpyyaml(f)
|
| 34 |
+
self.frontend = CosyVoiceFrontEnd(
|
| 35 |
+
configs["feat_extractor"],
|
| 36 |
+
"{}/campplus.onnx".format(model_dir),
|
| 37 |
+
"{}/speech_tokenizer_v1.onnx".format(model_dir),
|
| 38 |
+
)
|
| 39 |
+
self.model = CosyVoiceModel(configs["flow"], configs["hift"])
|
| 40 |
+
self.model.load(
|
| 41 |
+
"{}/flow.pt".format(model_dir),
|
| 42 |
+
"{}/hift.pt".format(model_dir),
|
| 43 |
+
)
|
| 44 |
+
self.model.flow = self.model.flow.to(torch.bfloat16)
|
| 45 |
+
del configs
|
| 46 |
+
|
| 47 |
+
def token_to_wav_offline(
|
| 48 |
+
self,
|
| 49 |
+
speech_token,
|
| 50 |
+
speech_feat,
|
| 51 |
+
speech_feat_len,
|
| 52 |
+
prompt_token,
|
| 53 |
+
prompt_token_len,
|
| 54 |
+
embedding,
|
| 55 |
+
):
|
| 56 |
+
tts_mel = self.model.flow.inference(
|
| 57 |
+
token=speech_token.to(self.model.device),
|
| 58 |
+
token_len=torch.tensor([speech_token.size(1)], dtype=torch.int32).to(
|
| 59 |
+
self.model.device
|
| 60 |
+
),
|
| 61 |
+
prompt_token=prompt_token.to(self.model.device),
|
| 62 |
+
prompt_token_len=prompt_token_len.to(self.model.device),
|
| 63 |
+
prompt_feat=speech_feat.to(self.model.device),
|
| 64 |
+
prompt_feat_len=speech_feat_len.to(self.model.device),
|
| 65 |
+
embedding=embedding.to(self.model.device),
|
| 66 |
+
)
|
| 67 |
+
tts_speech = self.model.hift.inference(mel=tts_mel.float())[0].cpu()
|
| 68 |
+
return tts_speech
|
cosyvoice/cli/frontend.py
ADDED
|
@@ -0,0 +1,106 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import onnxruntime
|
| 15 |
+
import torch
|
| 16 |
+
import numpy as np
|
| 17 |
+
import whisper
|
| 18 |
+
from typing import Callable
|
| 19 |
+
import torchaudio.compliance.kaldi as kaldi
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
class CosyVoiceFrontEnd:
|
| 23 |
+
|
| 24 |
+
def __init__(
|
| 25 |
+
self,
|
| 26 |
+
feat_extractor: Callable,
|
| 27 |
+
campplus_model: str,
|
| 28 |
+
speech_tokenizer_model: str,
|
| 29 |
+
):
|
| 30 |
+
self.feat_extractor = feat_extractor
|
| 31 |
+
self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
| 32 |
+
option = onnxruntime.SessionOptions()
|
| 33 |
+
option.graph_optimization_level = (
|
| 34 |
+
onnxruntime.GraphOptimizationLevel.ORT_ENABLE_ALL
|
| 35 |
+
)
|
| 36 |
+
option.intra_op_num_threads = 1
|
| 37 |
+
self.campplus_session = onnxruntime.InferenceSession(
|
| 38 |
+
campplus_model, sess_options=option, providers=["CPUExecutionProvider"]
|
| 39 |
+
)
|
| 40 |
+
self.speech_tokenizer_session = onnxruntime.InferenceSession(
|
| 41 |
+
speech_tokenizer_model,
|
| 42 |
+
sess_options=option,
|
| 43 |
+
providers=[
|
| 44 |
+
(
|
| 45 |
+
"CUDAExecutionProvider"
|
| 46 |
+
if torch.cuda.is_available()
|
| 47 |
+
else "CPUExecutionProvider"
|
| 48 |
+
)
|
| 49 |
+
],
|
| 50 |
+
)
|
| 51 |
+
|
| 52 |
+
def _extract_speech_token(self, speech):
|
| 53 |
+
assert (
|
| 54 |
+
speech.shape[1] / 16000 <= 30
|
| 55 |
+
), "do not support extract speech token for audio longer than 30s"
|
| 56 |
+
feat = whisper.log_mel_spectrogram(speech, n_mels=128)
|
| 57 |
+
speech_token = (
|
| 58 |
+
self.speech_tokenizer_session.run(
|
| 59 |
+
None,
|
| 60 |
+
{
|
| 61 |
+
self.speech_tokenizer_session.get_inputs()[0]
|
| 62 |
+
.name: feat.detach()
|
| 63 |
+
.cpu()
|
| 64 |
+
.numpy(),
|
| 65 |
+
self.speech_tokenizer_session.get_inputs()[1].name: np.array(
|
| 66 |
+
[feat.shape[2]], dtype=np.int32
|
| 67 |
+
),
|
| 68 |
+
},
|
| 69 |
+
)[0]
|
| 70 |
+
.flatten()
|
| 71 |
+
.tolist()
|
| 72 |
+
)
|
| 73 |
+
speech_token = torch.tensor([speech_token], dtype=torch.int32).to(self.device)
|
| 74 |
+
speech_token_len = torch.tensor([speech_token.shape[1]], dtype=torch.int32).to(
|
| 75 |
+
self.device
|
| 76 |
+
)
|
| 77 |
+
return speech_token, speech_token_len
|
| 78 |
+
|
| 79 |
+
def _extract_spk_embedding(self, speech):
|
| 80 |
+
feat = kaldi.fbank(speech, num_mel_bins=80, dither=0, sample_frequency=16000)
|
| 81 |
+
feat = feat - feat.mean(dim=0, keepdim=True)
|
| 82 |
+
embedding = (
|
| 83 |
+
self.campplus_session.run(
|
| 84 |
+
None,
|
| 85 |
+
{
|
| 86 |
+
self.campplus_session.get_inputs()[0]
|
| 87 |
+
.name: feat.unsqueeze(dim=0)
|
| 88 |
+
.cpu()
|
| 89 |
+
.numpy()
|
| 90 |
+
},
|
| 91 |
+
)[0]
|
| 92 |
+
.flatten()
|
| 93 |
+
.tolist()
|
| 94 |
+
)
|
| 95 |
+
embedding = torch.tensor([embedding]).to(self.device)
|
| 96 |
+
return embedding
|
| 97 |
+
|
| 98 |
+
def _extract_speech_feat(self, speech):
|
| 99 |
+
speech_feat = (
|
| 100 |
+
self.feat_extractor(speech).squeeze(dim=0).transpose(0, 1).to(self.device)
|
| 101 |
+
)
|
| 102 |
+
speech_feat = speech_feat.unsqueeze(dim=0)
|
| 103 |
+
speech_feat_len = torch.tensor([speech_feat.shape[1]], dtype=torch.int32).to(
|
| 104 |
+
self.device
|
| 105 |
+
)
|
| 106 |
+
return speech_feat, speech_feat_len
|
cosyvoice/cli/model.py
ADDED
|
@@ -0,0 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import torch
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
class CosyVoiceModel:
|
| 18 |
+
|
| 19 |
+
def __init__(
|
| 20 |
+
self,
|
| 21 |
+
flow: torch.nn.Module,
|
| 22 |
+
hift: torch.nn.Module,
|
| 23 |
+
):
|
| 24 |
+
self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
| 25 |
+
self.flow = flow
|
| 26 |
+
self.hift = hift
|
| 27 |
+
|
| 28 |
+
def load(self, flow_model, hift_model):
|
| 29 |
+
self.flow.load_state_dict(torch.load(flow_model, map_location=self.device))
|
| 30 |
+
self.flow.to(self.device).eval()
|
| 31 |
+
self.hift.load_state_dict(torch.load(hift_model, map_location=self.device))
|
| 32 |
+
self.hift.to(self.device).eval()
|
cosyvoice/flow/decoder.py
ADDED
|
@@ -0,0 +1,238 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import torch
|
| 15 |
+
import torch.nn as nn
|
| 16 |
+
from einops import pack, rearrange, repeat
|
| 17 |
+
from cosyvoice.matcha.decoder import (
|
| 18 |
+
SinusoidalPosEmb,
|
| 19 |
+
Block1D,
|
| 20 |
+
ResnetBlock1D,
|
| 21 |
+
Downsample1D,
|
| 22 |
+
TimestepEmbedding,
|
| 23 |
+
Upsample1D,
|
| 24 |
+
)
|
| 25 |
+
from cosyvoice.matcha.transformer import BasicTransformerBlock
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
class ConditionalDecoder(nn.Module):
|
| 29 |
+
def __init__(
|
| 30 |
+
self,
|
| 31 |
+
in_channels,
|
| 32 |
+
out_channels,
|
| 33 |
+
channels=(256, 256),
|
| 34 |
+
dropout=0.05,
|
| 35 |
+
attention_head_dim=64,
|
| 36 |
+
n_blocks=1,
|
| 37 |
+
num_mid_blocks=2,
|
| 38 |
+
num_heads=4,
|
| 39 |
+
act_fn="snake",
|
| 40 |
+
):
|
| 41 |
+
"""
|
| 42 |
+
This decoder requires an input with the same shape of the target. So, if your text content
|
| 43 |
+
is shorter or longer than the outputs, please re-sampling it before feeding to the decoder.
|
| 44 |
+
"""
|
| 45 |
+
super().__init__()
|
| 46 |
+
channels = tuple(channels)
|
| 47 |
+
self.in_channels = in_channels
|
| 48 |
+
self.out_channels = out_channels
|
| 49 |
+
|
| 50 |
+
self.time_embeddings = SinusoidalPosEmb(in_channels)
|
| 51 |
+
time_embed_dim = channels[0] * 4
|
| 52 |
+
self.time_mlp = TimestepEmbedding(
|
| 53 |
+
in_channels=in_channels,
|
| 54 |
+
time_embed_dim=time_embed_dim,
|
| 55 |
+
act_fn="silu",
|
| 56 |
+
)
|
| 57 |
+
self.down_blocks = nn.ModuleList([])
|
| 58 |
+
self.mid_blocks = nn.ModuleList([])
|
| 59 |
+
self.up_blocks = nn.ModuleList([])
|
| 60 |
+
|
| 61 |
+
output_channel = in_channels
|
| 62 |
+
for i in range(len(channels)): # pylint: disable=consider-using-enumerate
|
| 63 |
+
input_channel = output_channel
|
| 64 |
+
output_channel = channels[i]
|
| 65 |
+
is_last = i == len(channels) - 1
|
| 66 |
+
resnet = ResnetBlock1D(
|
| 67 |
+
dim=input_channel, dim_out=output_channel, time_emb_dim=time_embed_dim
|
| 68 |
+
)
|
| 69 |
+
transformer_blocks = nn.ModuleList(
|
| 70 |
+
[
|
| 71 |
+
BasicTransformerBlock(
|
| 72 |
+
dim=output_channel,
|
| 73 |
+
num_attention_heads=num_heads,
|
| 74 |
+
attention_head_dim=attention_head_dim,
|
| 75 |
+
dropout=dropout,
|
| 76 |
+
activation_fn=act_fn,
|
| 77 |
+
)
|
| 78 |
+
for _ in range(n_blocks)
|
| 79 |
+
]
|
| 80 |
+
)
|
| 81 |
+
downsample = (
|
| 82 |
+
Downsample1D(output_channel)
|
| 83 |
+
if not is_last
|
| 84 |
+
else nn.Conv1d(output_channel, output_channel, 3, padding=1)
|
| 85 |
+
)
|
| 86 |
+
self.down_blocks.append(
|
| 87 |
+
nn.ModuleList([resnet, transformer_blocks, downsample])
|
| 88 |
+
)
|
| 89 |
+
|
| 90 |
+
for _ in range(num_mid_blocks):
|
| 91 |
+
input_channel = channels[-1]
|
| 92 |
+
out_channels = channels[-1]
|
| 93 |
+
resnet = ResnetBlock1D(
|
| 94 |
+
dim=input_channel, dim_out=output_channel, time_emb_dim=time_embed_dim
|
| 95 |
+
)
|
| 96 |
+
|
| 97 |
+
transformer_blocks = nn.ModuleList(
|
| 98 |
+
[
|
| 99 |
+
BasicTransformerBlock(
|
| 100 |
+
dim=output_channel,
|
| 101 |
+
num_attention_heads=num_heads,
|
| 102 |
+
attention_head_dim=attention_head_dim,
|
| 103 |
+
dropout=dropout,
|
| 104 |
+
activation_fn=act_fn,
|
| 105 |
+
)
|
| 106 |
+
for _ in range(n_blocks)
|
| 107 |
+
]
|
| 108 |
+
)
|
| 109 |
+
|
| 110 |
+
self.mid_blocks.append(nn.ModuleList([resnet, transformer_blocks]))
|
| 111 |
+
|
| 112 |
+
channels = channels[::-1] + (channels[0],)
|
| 113 |
+
for i in range(len(channels) - 1):
|
| 114 |
+
input_channel = channels[i] * 2
|
| 115 |
+
output_channel = channels[i + 1]
|
| 116 |
+
is_last = i == len(channels) - 2
|
| 117 |
+
resnet = ResnetBlock1D(
|
| 118 |
+
dim=input_channel,
|
| 119 |
+
dim_out=output_channel,
|
| 120 |
+
time_emb_dim=time_embed_dim,
|
| 121 |
+
)
|
| 122 |
+
transformer_blocks = nn.ModuleList(
|
| 123 |
+
[
|
| 124 |
+
BasicTransformerBlock(
|
| 125 |
+
dim=output_channel,
|
| 126 |
+
num_attention_heads=num_heads,
|
| 127 |
+
attention_head_dim=attention_head_dim,
|
| 128 |
+
dropout=dropout,
|
| 129 |
+
activation_fn=act_fn,
|
| 130 |
+
)
|
| 131 |
+
for _ in range(n_blocks)
|
| 132 |
+
]
|
| 133 |
+
)
|
| 134 |
+
upsample = (
|
| 135 |
+
Upsample1D(output_channel, use_conv_transpose=True)
|
| 136 |
+
if not is_last
|
| 137 |
+
else nn.Conv1d(output_channel, output_channel, 3, padding=1)
|
| 138 |
+
)
|
| 139 |
+
self.up_blocks.append(nn.ModuleList([resnet, transformer_blocks, upsample]))
|
| 140 |
+
self.final_block = Block1D(channels[-1], channels[-1])
|
| 141 |
+
self.final_proj = nn.Conv1d(channels[-1], self.out_channels, 1)
|
| 142 |
+
self.initialize_weights()
|
| 143 |
+
|
| 144 |
+
def initialize_weights(self):
|
| 145 |
+
for m in self.modules():
|
| 146 |
+
if isinstance(m, nn.Conv1d):
|
| 147 |
+
nn.init.kaiming_normal_(m.weight, nonlinearity="relu")
|
| 148 |
+
if m.bias is not None:
|
| 149 |
+
nn.init.constant_(m.bias, 0)
|
| 150 |
+
elif isinstance(m, nn.GroupNorm):
|
| 151 |
+
nn.init.constant_(m.weight, 1)
|
| 152 |
+
nn.init.constant_(m.bias, 0)
|
| 153 |
+
elif isinstance(m, nn.Linear):
|
| 154 |
+
nn.init.kaiming_normal_(m.weight, nonlinearity="relu")
|
| 155 |
+
if m.bias is not None:
|
| 156 |
+
nn.init.constant_(m.bias, 0)
|
| 157 |
+
|
| 158 |
+
def forward(self, x, mask, mu, t, spks=None, cond=None):
|
| 159 |
+
"""Forward pass of the UNet1DConditional model.
|
| 160 |
+
|
| 161 |
+
Args:
|
| 162 |
+
x (torch.Tensor): shape (batch_size, in_channels, time)
|
| 163 |
+
mask (_type_): shape (batch_size, 1, time)
|
| 164 |
+
t (_type_): shape (batch_size)
|
| 165 |
+
spks (_type_, optional): shape: (batch_size, condition_channels). Defaults to None.
|
| 166 |
+
cond (_type_, optional): placeholder for future use. Defaults to None.
|
| 167 |
+
|
| 168 |
+
Raises:
|
| 169 |
+
ValueError: _description_
|
| 170 |
+
ValueError: _description_
|
| 171 |
+
|
| 172 |
+
Returns:
|
| 173 |
+
_type_: _description_
|
| 174 |
+
"""
|
| 175 |
+
|
| 176 |
+
t = self.time_embeddings(t).to(t.dtype)
|
| 177 |
+
t = self.time_mlp(t)
|
| 178 |
+
|
| 179 |
+
x = pack([x, mu], "b * t")[0]
|
| 180 |
+
|
| 181 |
+
if spks is not None:
|
| 182 |
+
spks = repeat(spks, "b c -> b c t", t=x.shape[-1])
|
| 183 |
+
x = pack([x, spks], "b * t")[0]
|
| 184 |
+
if cond is not None:
|
| 185 |
+
x = pack([x, cond], "b * t")[0]
|
| 186 |
+
|
| 187 |
+
hiddens = []
|
| 188 |
+
masks = [mask]
|
| 189 |
+
for resnet, transformer_blocks, downsample in self.down_blocks:
|
| 190 |
+
mask_down = masks[-1]
|
| 191 |
+
x = resnet(
|
| 192 |
+
x.to(torch.bfloat16), mask_down.to(torch.bfloat16), t.to(torch.bfloat16)
|
| 193 |
+
)
|
| 194 |
+
x = rearrange(x, "b c t -> b t c").contiguous()
|
| 195 |
+
# attn_mask = torch.matmul(mask_down.transpose(1, 2).contiguous(), mask_down)
|
| 196 |
+
for transformer_block in transformer_blocks:
|
| 197 |
+
x = transformer_block(
|
| 198 |
+
hidden_states=x,
|
| 199 |
+
# attention_mask=attn_mask,
|
| 200 |
+
timestep=t,
|
| 201 |
+
)
|
| 202 |
+
x = rearrange(x, "b t c -> b c t").contiguous()
|
| 203 |
+
hiddens.append(x) # Save hidden states for skip connections
|
| 204 |
+
x = downsample(x * mask_down)
|
| 205 |
+
masks.append(mask_down[:, :, ::2])
|
| 206 |
+
masks = masks[:-1]
|
| 207 |
+
mask_mid = masks[-1]
|
| 208 |
+
|
| 209 |
+
for resnet, transformer_blocks in self.mid_blocks:
|
| 210 |
+
x = resnet(x, mask_mid, t)
|
| 211 |
+
x = rearrange(x, "b c t -> b t c").contiguous()
|
| 212 |
+
# attn_mask = torch.matmul(mask_mid.transpose(1, 2).contiguous(), mask_mid)
|
| 213 |
+
for transformer_block in transformer_blocks:
|
| 214 |
+
x = transformer_block(
|
| 215 |
+
hidden_states=x,
|
| 216 |
+
# attention_mask=attn_mask,
|
| 217 |
+
timestep=t,
|
| 218 |
+
)
|
| 219 |
+
x = rearrange(x, "b t c -> b c t").contiguous()
|
| 220 |
+
|
| 221 |
+
for resnet, transformer_blocks, upsample in self.up_blocks:
|
| 222 |
+
mask_up = masks.pop()
|
| 223 |
+
skip = hiddens.pop()
|
| 224 |
+
x = pack([x[:, :, : skip.shape[-1]], skip], "b * t")[0]
|
| 225 |
+
x = resnet(x, mask_up, t)
|
| 226 |
+
x = rearrange(x, "b c t -> b t c").contiguous()
|
| 227 |
+
# attn_mask = torch.matmul(mask_up.transpose(1, 2).contiguous(), mask_up)
|
| 228 |
+
for transformer_block in transformer_blocks:
|
| 229 |
+
x = transformer_block(
|
| 230 |
+
hidden_states=x,
|
| 231 |
+
# attention_mask=attn_mask,
|
| 232 |
+
timestep=t,
|
| 233 |
+
)
|
| 234 |
+
x = rearrange(x, "b t c -> b c t").contiguous()
|
| 235 |
+
x = upsample(x * mask_up)
|
| 236 |
+
x = self.final_block(x, mask_up)
|
| 237 |
+
output = self.final_proj(x * mask_up)
|
| 238 |
+
return output * mask
|
cosyvoice/flow/flow.py
ADDED
|
@@ -0,0 +1,196 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import logging
|
| 15 |
+
import random
|
| 16 |
+
from typing import Dict, Optional
|
| 17 |
+
import torch
|
| 18 |
+
import torch.nn as nn
|
| 19 |
+
from torch.nn import functional as F
|
| 20 |
+
from omegaconf import DictConfig
|
| 21 |
+
from cosyvoice.utils.mask import make_pad_mask
|
| 22 |
+
import time
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
class MaskedDiffWithXvec(torch.nn.Module):
|
| 26 |
+
def __init__(
|
| 27 |
+
self,
|
| 28 |
+
input_size: int = 512,
|
| 29 |
+
output_size: int = 80,
|
| 30 |
+
spk_embed_dim: int = 192,
|
| 31 |
+
output_type: str = "mel",
|
| 32 |
+
vocab_size: int = 4096,
|
| 33 |
+
input_frame_rate: int = 50,
|
| 34 |
+
only_mask_loss: bool = True,
|
| 35 |
+
encoder: torch.nn.Module = None,
|
| 36 |
+
length_regulator: torch.nn.Module = None,
|
| 37 |
+
decoder: torch.nn.Module = None,
|
| 38 |
+
decoder_conf: Dict = {
|
| 39 |
+
"in_channels": 240,
|
| 40 |
+
"out_channel": 80,
|
| 41 |
+
"spk_emb_dim": 80,
|
| 42 |
+
"n_spks": 1,
|
| 43 |
+
"cfm_params": DictConfig(
|
| 44 |
+
{
|
| 45 |
+
"sigma_min": 1e-06,
|
| 46 |
+
"solver": "euler",
|
| 47 |
+
"t_scheduler": "cosine",
|
| 48 |
+
"training_cfg_rate": 0.2,
|
| 49 |
+
"inference_cfg_rate": 0.7,
|
| 50 |
+
"reg_loss_type": "l1",
|
| 51 |
+
}
|
| 52 |
+
),
|
| 53 |
+
"decoder_params": {
|
| 54 |
+
"channels": [256, 256],
|
| 55 |
+
"dropout": 0.0,
|
| 56 |
+
"attention_head_dim": 64,
|
| 57 |
+
"n_blocks": 4,
|
| 58 |
+
"num_mid_blocks": 12,
|
| 59 |
+
"num_heads": 8,
|
| 60 |
+
"act_fn": "gelu",
|
| 61 |
+
},
|
| 62 |
+
},
|
| 63 |
+
mel_feat_conf: Dict = {
|
| 64 |
+
"n_fft": 1024,
|
| 65 |
+
"num_mels": 80,
|
| 66 |
+
"sampling_rate": 22050,
|
| 67 |
+
"hop_size": 256,
|
| 68 |
+
"win_size": 1024,
|
| 69 |
+
"fmin": 0,
|
| 70 |
+
"fmax": 8000,
|
| 71 |
+
},
|
| 72 |
+
):
|
| 73 |
+
super().__init__()
|
| 74 |
+
self.input_size = input_size
|
| 75 |
+
self.output_size = output_size
|
| 76 |
+
self.decoder_conf = decoder_conf
|
| 77 |
+
self.mel_feat_conf = mel_feat_conf
|
| 78 |
+
self.vocab_size = vocab_size
|
| 79 |
+
self.output_type = output_type
|
| 80 |
+
self.input_frame_rate = input_frame_rate
|
| 81 |
+
logging.info(f"input frame rate={self.input_frame_rate}")
|
| 82 |
+
self.input_embedding = nn.Embedding(vocab_size, input_size)
|
| 83 |
+
self.spk_embed_affine_layer = torch.nn.Linear(spk_embed_dim, output_size)
|
| 84 |
+
self.encoder = encoder
|
| 85 |
+
self.encoder_proj = torch.nn.Linear(self.encoder.output_size(), output_size)
|
| 86 |
+
self.decoder = decoder
|
| 87 |
+
self.length_regulator = length_regulator
|
| 88 |
+
self.only_mask_loss = only_mask_loss
|
| 89 |
+
|
| 90 |
+
def forward(
|
| 91 |
+
self,
|
| 92 |
+
batch: dict,
|
| 93 |
+
device: torch.device,
|
| 94 |
+
) -> Dict[str, Optional[torch.Tensor]]:
|
| 95 |
+
token = batch["speech_token"].to(device)
|
| 96 |
+
token_len = batch["speech_token_len"].to(device)
|
| 97 |
+
feat = batch["speech_feat"].to(device)
|
| 98 |
+
feat_len = batch["speech_feat_len"].to(device)
|
| 99 |
+
embedding = batch["embedding"].to(device)
|
| 100 |
+
|
| 101 |
+
# xvec projection
|
| 102 |
+
embedding = F.normalize(embedding, dim=1)
|
| 103 |
+
embedding = self.spk_embed_affine_layer(embedding)
|
| 104 |
+
|
| 105 |
+
# concat text and prompt_text
|
| 106 |
+
mask = (~make_pad_mask(token_len)).float().unsqueeze(-1).to(device)
|
| 107 |
+
token = self.input_embedding(torch.clamp(token, min=0)) * mask
|
| 108 |
+
|
| 109 |
+
# text encode
|
| 110 |
+
h, h_lengths = self.encoder(token, token_len)
|
| 111 |
+
h = self.encoder_proj(h)
|
| 112 |
+
h, h_lengths = self.length_regulator(h, feat_len)
|
| 113 |
+
|
| 114 |
+
# get conditions
|
| 115 |
+
conds = torch.zeros(feat.shape, device=token.device)
|
| 116 |
+
for i, j in enumerate(feat_len):
|
| 117 |
+
if random.random() < 0.5:
|
| 118 |
+
continue
|
| 119 |
+
index = random.randint(0, int(0.3 * j))
|
| 120 |
+
conds[i, :index] = feat[i, :index]
|
| 121 |
+
conds = conds.transpose(1, 2)
|
| 122 |
+
|
| 123 |
+
mask = (~make_pad_mask(feat_len)).to(h)
|
| 124 |
+
feat = F.interpolate(
|
| 125 |
+
feat.unsqueeze(dim=1), size=h.shape[1:], mode="nearest"
|
| 126 |
+
).squeeze(dim=1)
|
| 127 |
+
loss, _ = self.decoder.compute_loss(
|
| 128 |
+
feat.transpose(1, 2).contiguous(),
|
| 129 |
+
mask.unsqueeze(1),
|
| 130 |
+
h.transpose(1, 2).contiguous(),
|
| 131 |
+
embedding,
|
| 132 |
+
cond=conds,
|
| 133 |
+
)
|
| 134 |
+
return {"loss": loss}
|
| 135 |
+
|
| 136 |
+
@torch.inference_mode()
|
| 137 |
+
def inference(
|
| 138 |
+
self,
|
| 139 |
+
token,
|
| 140 |
+
token_len,
|
| 141 |
+
prompt_token,
|
| 142 |
+
prompt_token_len,
|
| 143 |
+
prompt_feat,
|
| 144 |
+
prompt_feat_len,
|
| 145 |
+
embedding,
|
| 146 |
+
):
|
| 147 |
+
assert token.shape[0] == 1
|
| 148 |
+
# xvec projection
|
| 149 |
+
embedding = F.normalize(embedding, dim=1)
|
| 150 |
+
embedding = self.spk_embed_affine_layer(embedding)
|
| 151 |
+
|
| 152 |
+
# concat text and prompt_text
|
| 153 |
+
token_len1, token_len2 = prompt_token.shape[1], token.shape[1]
|
| 154 |
+
# text encode
|
| 155 |
+
token, token_len = (
|
| 156 |
+
torch.concat([prompt_token, token], dim=1),
|
| 157 |
+
prompt_token_len + token_len,
|
| 158 |
+
)
|
| 159 |
+
token = self.input_embedding(torch.clamp(token, min=0))
|
| 160 |
+
h, _ = self.encoder.inference(token, token_len)
|
| 161 |
+
h = self.encoder_proj(h)
|
| 162 |
+
mel_len1, mel_len2 = prompt_feat.shape[1], int(
|
| 163 |
+
token_len2
|
| 164 |
+
/ self.input_frame_rate
|
| 165 |
+
* self.mel_feat_conf["sampling_rate"]
|
| 166 |
+
/ self.mel_feat_conf["hop_size"]
|
| 167 |
+
)
|
| 168 |
+
|
| 169 |
+
h, _ = self.length_regulator.inference(
|
| 170 |
+
h[:, :token_len1],
|
| 171 |
+
h[:, token_len1:],
|
| 172 |
+
mel_len1,
|
| 173 |
+
mel_len2,
|
| 174 |
+
)
|
| 175 |
+
|
| 176 |
+
# get conditions
|
| 177 |
+
conds = torch.zeros(
|
| 178 |
+
[1, mel_len1 + mel_len2, self.output_size], device=token.device
|
| 179 |
+
)
|
| 180 |
+
conds[:, :mel_len1] = prompt_feat
|
| 181 |
+
conds = conds.transpose(1, 2)
|
| 182 |
+
|
| 183 |
+
# mask = (~make_pad_mask(torch.tensor([mel_len1 + mel_len2]))).to(h)
|
| 184 |
+
mask = torch.ones(
|
| 185 |
+
[1, mel_len1 + mel_len2], device=h.device, dtype=torch.bfloat16
|
| 186 |
+
)
|
| 187 |
+
feat = self.decoder(
|
| 188 |
+
mu=h.transpose(1, 2).contiguous(),
|
| 189 |
+
mask=mask.unsqueeze(1),
|
| 190 |
+
spks=embedding,
|
| 191 |
+
cond=conds,
|
| 192 |
+
n_timesteps=10,
|
| 193 |
+
)
|
| 194 |
+
feat = feat[:, :, mel_len1:]
|
| 195 |
+
assert feat.shape[2] == mel_len2
|
| 196 |
+
return feat
|
cosyvoice/flow/flow_matching.py
ADDED
|
@@ -0,0 +1,315 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import time
|
| 15 |
+
import torch
|
| 16 |
+
import torch.nn.functional as F
|
| 17 |
+
from cosyvoice.matcha.flow_matching import BASECFM
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
class ConditionalCFM(BASECFM):
|
| 21 |
+
def __init__(
|
| 22 |
+
self,
|
| 23 |
+
in_channels,
|
| 24 |
+
cfm_params,
|
| 25 |
+
n_spks=1,
|
| 26 |
+
spk_emb_dim=64,
|
| 27 |
+
estimator: torch.nn.Module = None,
|
| 28 |
+
):
|
| 29 |
+
super().__init__(
|
| 30 |
+
n_feats=in_channels,
|
| 31 |
+
cfm_params=cfm_params,
|
| 32 |
+
n_spks=n_spks,
|
| 33 |
+
spk_emb_dim=spk_emb_dim,
|
| 34 |
+
)
|
| 35 |
+
self.t_scheduler = cfm_params.t_scheduler
|
| 36 |
+
self.training_cfg_rate = cfm_params.training_cfg_rate
|
| 37 |
+
self.inference_cfg_rate = cfm_params.inference_cfg_rate
|
| 38 |
+
in_channels = in_channels + (spk_emb_dim if n_spks > 0 else 0)
|
| 39 |
+
# Just change the architecture of the estimator here
|
| 40 |
+
self.estimator = estimator
|
| 41 |
+
self.inference_graphs = {}
|
| 42 |
+
self.inference_buffers = {}
|
| 43 |
+
# self.capture_inference()
|
| 44 |
+
|
| 45 |
+
@torch.inference_mode()
|
| 46 |
+
def forward(
|
| 47 |
+
self,
|
| 48 |
+
mu,
|
| 49 |
+
mask,
|
| 50 |
+
n_timesteps,
|
| 51 |
+
temperature=1.0,
|
| 52 |
+
spks=None,
|
| 53 |
+
cond=None,
|
| 54 |
+
):
|
| 55 |
+
"""Forward diffusion
|
| 56 |
+
|
| 57 |
+
Args:
|
| 58 |
+
mu (torch.Tensor): output of encoder
|
| 59 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 60 |
+
mask (torch.Tensor): output_mask
|
| 61 |
+
shape: (batch_size, 1, mel_timesteps)
|
| 62 |
+
n_timesteps (int): number of diffusion steps
|
| 63 |
+
temperature (float, optional): temperature for scaling noise. Defaults to 1.0.
|
| 64 |
+
spks (torch.Tensor, optional): speaker ids. Defaults to None.
|
| 65 |
+
shape: (batch_size, spk_emb_dim)
|
| 66 |
+
cond: Not used but kept for future purposes
|
| 67 |
+
|
| 68 |
+
Returns:
|
| 69 |
+
sample: generated mel-spectrogram
|
| 70 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 71 |
+
"""
|
| 72 |
+
z = torch.randn_like(mu) * temperature
|
| 73 |
+
t_span = torch.linspace(0, 1, n_timesteps + 1, device=mu.device, dtype=mu.dtype)
|
| 74 |
+
if self.t_scheduler == "cosine":
|
| 75 |
+
t_span = 1 - torch.cos(t_span * 0.5 * torch.pi)
|
| 76 |
+
return self.solve_euler(
|
| 77 |
+
z, t_span=t_span, mu=mu, mask=mask, spks=spks, cond=cond
|
| 78 |
+
)
|
| 79 |
+
|
| 80 |
+
@torch.inference_mode()
|
| 81 |
+
def capture_inference(self, seq_len_to_capture=list(range(128, 512, 8))):
|
| 82 |
+
start_time = time.time()
|
| 83 |
+
print(
|
| 84 |
+
f"capture_inference for ConditionalCFM solve euler, seq_len_to_capture: {seq_len_to_capture}"
|
| 85 |
+
)
|
| 86 |
+
for seq_len in seq_len_to_capture:
|
| 87 |
+
static_z = torch.randn(
|
| 88 |
+
1, 80, seq_len, device=torch.device("cuda"), dtype=torch.bfloat16
|
| 89 |
+
)
|
| 90 |
+
static_t_span = torch.linspace(
|
| 91 |
+
0, 1, 11, device=torch.device("cuda"), dtype=torch.bfloat16
|
| 92 |
+
) # only capture at 10 steps
|
| 93 |
+
static_mu = torch.randn(
|
| 94 |
+
1, 80, seq_len, device=torch.device("cuda"), dtype=torch.bfloat16
|
| 95 |
+
)
|
| 96 |
+
static_mask = torch.ones(
|
| 97 |
+
1, 1, seq_len, device=torch.device("cuda"), dtype=torch.bfloat16
|
| 98 |
+
)
|
| 99 |
+
static_spks = torch.randn(
|
| 100 |
+
1, 80, device=torch.device("cuda"), dtype=torch.bfloat16
|
| 101 |
+
)
|
| 102 |
+
static_cond = torch.randn(
|
| 103 |
+
1, 80, seq_len, device=torch.device("cuda"), dtype=torch.float32
|
| 104 |
+
)
|
| 105 |
+
static_out = torch.randn(
|
| 106 |
+
1, 80, seq_len, device=torch.device("cuda"), dtype=torch.bfloat16
|
| 107 |
+
)
|
| 108 |
+
|
| 109 |
+
self._solve_euler_impl(
|
| 110 |
+
static_z,
|
| 111 |
+
t_span=static_t_span,
|
| 112 |
+
mu=static_mu,
|
| 113 |
+
mask=static_mask,
|
| 114 |
+
spks=static_spks,
|
| 115 |
+
cond=static_cond,
|
| 116 |
+
)
|
| 117 |
+
torch.cuda.synchronize()
|
| 118 |
+
|
| 119 |
+
g = torch.cuda.CUDAGraph()
|
| 120 |
+
with torch.cuda.graph(g):
|
| 121 |
+
static_out = self._solve_euler_impl(
|
| 122 |
+
static_z,
|
| 123 |
+
t_span=static_t_span,
|
| 124 |
+
mu=static_mu,
|
| 125 |
+
mask=static_mask,
|
| 126 |
+
spks=static_spks,
|
| 127 |
+
cond=static_cond,
|
| 128 |
+
)
|
| 129 |
+
|
| 130 |
+
self.inference_buffers[seq_len] = {
|
| 131 |
+
"z": static_z,
|
| 132 |
+
"t_span": static_t_span,
|
| 133 |
+
"mu": static_mu,
|
| 134 |
+
"mask": static_mask,
|
| 135 |
+
"spks": static_spks,
|
| 136 |
+
"cond": static_cond,
|
| 137 |
+
"out": static_out,
|
| 138 |
+
}
|
| 139 |
+
self.inference_graphs[seq_len] = g
|
| 140 |
+
end_time = time.time()
|
| 141 |
+
print(
|
| 142 |
+
f"capture_inference for ConditionalCFM solve euler, time elapsed: {end_time - start_time}"
|
| 143 |
+
)
|
| 144 |
+
|
| 145 |
+
def solve_euler(self, x, t_span, mu, mask, spks, cond):
|
| 146 |
+
if hasattr(self, "inference_graphs") and len(self.inference_graphs) > 0:
|
| 147 |
+
curr_seq_len = x.shape[2]
|
| 148 |
+
|
| 149 |
+
available_lengths = sorted(list(self.inference_graphs.keys()))
|
| 150 |
+
|
| 151 |
+
if curr_seq_len <= max(available_lengths):
|
| 152 |
+
target_len = min(available_lengths, key=lambda x: abs(x - curr_seq_len))
|
| 153 |
+
if target_len == curr_seq_len:
|
| 154 |
+
padded_x = x
|
| 155 |
+
padded_mu = mu
|
| 156 |
+
padded_mask = mask
|
| 157 |
+
if cond is not None:
|
| 158 |
+
padded_cond = cond
|
| 159 |
+
else:
|
| 160 |
+
padded_x = torch.randn(
|
| 161 |
+
(x.shape[0], x.shape[1], target_len),
|
| 162 |
+
dtype=x.dtype,
|
| 163 |
+
device=x.device,
|
| 164 |
+
)
|
| 165 |
+
padded_x[:, :, :curr_seq_len] = x
|
| 166 |
+
|
| 167 |
+
padded_mu = torch.randn(
|
| 168 |
+
(mu.shape[0], mu.shape[1], target_len),
|
| 169 |
+
dtype=mu.dtype,
|
| 170 |
+
device=mu.device,
|
| 171 |
+
)
|
| 172 |
+
padded_mu[:, :, :curr_seq_len] = mu
|
| 173 |
+
|
| 174 |
+
# FIXME(ys): uses zeros and maskgroupnorm
|
| 175 |
+
padded_mask = torch.ones(
|
| 176 |
+
(mask.shape[0], mask.shape[1], target_len),
|
| 177 |
+
dtype=mask.dtype,
|
| 178 |
+
device=mask.device,
|
| 179 |
+
)
|
| 180 |
+
|
| 181 |
+
if cond is not None:
|
| 182 |
+
padded_cond = torch.randn(
|
| 183 |
+
(cond.shape[0], cond.shape[1], target_len),
|
| 184 |
+
dtype=cond.dtype,
|
| 185 |
+
device=cond.device,
|
| 186 |
+
)
|
| 187 |
+
padded_cond[:, :, :curr_seq_len] = cond
|
| 188 |
+
|
| 189 |
+
buffer = self.inference_buffers[target_len]
|
| 190 |
+
buffer["z"].copy_(padded_x)
|
| 191 |
+
buffer["t_span"].copy_(t_span)
|
| 192 |
+
buffer["mu"].copy_(padded_mu)
|
| 193 |
+
buffer["mask"].copy_(padded_mask)
|
| 194 |
+
buffer["spks"].copy_(spks)
|
| 195 |
+
if cond is not None:
|
| 196 |
+
buffer["cond"].copy_(padded_cond)
|
| 197 |
+
|
| 198 |
+
self.inference_graphs[target_len].replay()
|
| 199 |
+
|
| 200 |
+
output = buffer["out"][:, :, :curr_seq_len]
|
| 201 |
+
return output
|
| 202 |
+
|
| 203 |
+
return self._solve_euler_impl(x, t_span, mu, mask, spks, cond)
|
| 204 |
+
|
| 205 |
+
def _solve_euler_impl(self, x, t_span, mu, mask, spks, cond):
|
| 206 |
+
"""
|
| 207 |
+
Fixed euler solver for ODEs.
|
| 208 |
+
Args:
|
| 209 |
+
x (torch.Tensor): random noise
|
| 210 |
+
t_span (torch.Tensor): n_timesteps interpolated
|
| 211 |
+
shape: (n_timesteps + 1,)
|
| 212 |
+
mu (torch.Tensor): output of encoder
|
| 213 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 214 |
+
mask (torch.Tensor): output_mask
|
| 215 |
+
shape: (batch_size, 1, mel_timesteps)
|
| 216 |
+
spks (torch.Tensor, optional): speaker ids. Defaults to None.
|
| 217 |
+
shape: (batch_size, spk_emb_dim)
|
| 218 |
+
cond: Not used but kept for future purposes
|
| 219 |
+
"""
|
| 220 |
+
t, _, dt = t_span[0], t_span[-1], t_span[1] - t_span[0]
|
| 221 |
+
t = t.unsqueeze(dim=0)
|
| 222 |
+
|
| 223 |
+
# I am storing this because I can later plot it by putting a debugger here and saving it to a file
|
| 224 |
+
# Or in future might add like a return_all_steps flag
|
| 225 |
+
sol = []
|
| 226 |
+
|
| 227 |
+
for step in range(1, len(t_span)):
|
| 228 |
+
if self.inference_cfg_rate > 0:
|
| 229 |
+
x_double = torch.cat([x, x], dim=0)
|
| 230 |
+
mask_double = torch.cat([mask, mask], dim=0)
|
| 231 |
+
mu_double = torch.cat([mu, torch.zeros_like(mu)], dim=0)
|
| 232 |
+
t_double = torch.cat([t, t], dim=0)
|
| 233 |
+
spks_double = (
|
| 234 |
+
torch.cat([spks, torch.zeros_like(spks)], dim=0)
|
| 235 |
+
if spks is not None
|
| 236 |
+
else None
|
| 237 |
+
)
|
| 238 |
+
cond_double = torch.cat([cond, torch.zeros_like(cond)], dim=0)
|
| 239 |
+
|
| 240 |
+
dphi_dt_double = self.forward_estimator(
|
| 241 |
+
x_double, mask_double, mu_double, t_double, spks_double, cond_double
|
| 242 |
+
)
|
| 243 |
+
|
| 244 |
+
dphi_dt, cfg_dphi_dt = torch.chunk(dphi_dt_double, 2, dim=0)
|
| 245 |
+
dphi_dt = (
|
| 246 |
+
1.0 + self.inference_cfg_rate
|
| 247 |
+
) * dphi_dt - self.inference_cfg_rate * cfg_dphi_dt
|
| 248 |
+
else:
|
| 249 |
+
dphi_dt = self.forward_estimator(x, mask, mu, t, spks, cond)
|
| 250 |
+
|
| 251 |
+
x = x + dt * dphi_dt
|
| 252 |
+
t = t + dt
|
| 253 |
+
sol.append(x)
|
| 254 |
+
if step < len(t_span) - 1:
|
| 255 |
+
dt = t_span[step + 1] - t
|
| 256 |
+
|
| 257 |
+
return sol[-1]
|
| 258 |
+
|
| 259 |
+
def forward_estimator(self, x, mask, mu, t, spks, cond):
|
| 260 |
+
if isinstance(self.estimator, torch.nn.Module):
|
| 261 |
+
return self.estimator.forward(x, mask, mu, t, spks, cond)
|
| 262 |
+
else:
|
| 263 |
+
ort_inputs = {
|
| 264 |
+
"x": x.cpu().numpy(),
|
| 265 |
+
"mask": mask.cpu().numpy(),
|
| 266 |
+
"mu": mu.cpu().numpy(),
|
| 267 |
+
"t": t.cpu().numpy(),
|
| 268 |
+
"spks": spks.cpu().numpy(),
|
| 269 |
+
"cond": cond.cpu().numpy(),
|
| 270 |
+
}
|
| 271 |
+
output = self.estimator.run(None, ort_inputs)[0]
|
| 272 |
+
return torch.tensor(output, dtype=x.dtype, device=x.device)
|
| 273 |
+
|
| 274 |
+
def compute_loss(self, x1, mask, mu, spks=None, cond=None):
|
| 275 |
+
"""Computes diffusion loss
|
| 276 |
+
|
| 277 |
+
Args:
|
| 278 |
+
x1 (torch.Tensor): Target
|
| 279 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 280 |
+
mask (torch.Tensor): target mask
|
| 281 |
+
shape: (batch_size, 1, mel_timesteps)
|
| 282 |
+
mu (torch.Tensor): output of encoder
|
| 283 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 284 |
+
spks (torch.Tensor, optional): speaker embedding. Defaults to None.
|
| 285 |
+
shape: (batch_size, spk_emb_dim)
|
| 286 |
+
|
| 287 |
+
Returns:
|
| 288 |
+
loss: conditional flow matching loss
|
| 289 |
+
y: conditional flow
|
| 290 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 291 |
+
"""
|
| 292 |
+
b, _, t = mu.shape
|
| 293 |
+
|
| 294 |
+
# random timestep
|
| 295 |
+
t = torch.rand([b, 1, 1], device=mu.device, dtype=mu.dtype)
|
| 296 |
+
if self.t_scheduler == "cosine":
|
| 297 |
+
t = 1 - torch.cos(t * 0.5 * torch.pi)
|
| 298 |
+
# sample noise p(x_0)
|
| 299 |
+
z = torch.randn_like(x1)
|
| 300 |
+
|
| 301 |
+
y = (1 - (1 - self.sigma_min) * t) * z + t * x1
|
| 302 |
+
u = x1 - (1 - self.sigma_min) * z
|
| 303 |
+
|
| 304 |
+
# during training, we randomly drop condition to trade off mode coverage and sample fidelity
|
| 305 |
+
if self.training_cfg_rate > 0:
|
| 306 |
+
cfg_mask = torch.rand(b, device=x1.device) > self.training_cfg_rate
|
| 307 |
+
mu = mu * cfg_mask.view(-1, 1, 1)
|
| 308 |
+
spks = spks * cfg_mask.view(-1, 1)
|
| 309 |
+
cond = cond * cfg_mask.view(-1, 1, 1)
|
| 310 |
+
|
| 311 |
+
pred = self.estimator(y, mask, mu, t.squeeze(), spks, cond)
|
| 312 |
+
loss = F.mse_loss(pred * mask, u * mask, reduction="sum") / (
|
| 313 |
+
torch.sum(mask) * u.shape[1]
|
| 314 |
+
)
|
| 315 |
+
return loss, y
|
cosyvoice/flow/length_regulator.py
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
from typing import Tuple
|
| 15 |
+
import torch.nn as nn
|
| 16 |
+
import torch
|
| 17 |
+
from torch.nn import functional as F
|
| 18 |
+
from cosyvoice.utils.mask import make_pad_mask
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
class InterpolateRegulator(nn.Module):
|
| 22 |
+
def __init__(
|
| 23 |
+
self,
|
| 24 |
+
channels: int,
|
| 25 |
+
sampling_ratios: Tuple,
|
| 26 |
+
out_channels: int = None,
|
| 27 |
+
groups: int = 1,
|
| 28 |
+
):
|
| 29 |
+
super().__init__()
|
| 30 |
+
self.sampling_ratios = sampling_ratios
|
| 31 |
+
out_channels = out_channels or channels
|
| 32 |
+
model = nn.ModuleList([])
|
| 33 |
+
if len(sampling_ratios) > 0:
|
| 34 |
+
for _ in sampling_ratios:
|
| 35 |
+
module = nn.Conv1d(channels, channels, 3, 1, 1)
|
| 36 |
+
norm = nn.GroupNorm(groups, channels)
|
| 37 |
+
act = nn.Mish()
|
| 38 |
+
model.extend([module, norm, act])
|
| 39 |
+
model.append(nn.Conv1d(channels, out_channels, 1, 1))
|
| 40 |
+
self.model = nn.Sequential(*model)
|
| 41 |
+
|
| 42 |
+
def forward(self, x, ylens=None):
|
| 43 |
+
# x in (B, T, D)
|
| 44 |
+
mask = (~make_pad_mask(ylens)).to(x).unsqueeze(-1)
|
| 45 |
+
x = F.interpolate(
|
| 46 |
+
x.transpose(1, 2).contiguous(), size=ylens.max(), mode="linear"
|
| 47 |
+
)
|
| 48 |
+
out = self.model(x).transpose(1, 2).contiguous()
|
| 49 |
+
olens = ylens
|
| 50 |
+
return out * mask, olens
|
| 51 |
+
|
| 52 |
+
def inference(self, x1, x2, mel_len1, mel_len2):
|
| 53 |
+
# x in (B, T, D)
|
| 54 |
+
x2 = F.interpolate(
|
| 55 |
+
x2.transpose(1, 2).contiguous(), size=mel_len2, mode="linear"
|
| 56 |
+
)
|
| 57 |
+
if x1.shape[1] != 0:
|
| 58 |
+
x1 = F.interpolate(
|
| 59 |
+
x1.transpose(1, 2).contiguous(), size=mel_len1, mode="linear"
|
| 60 |
+
)
|
| 61 |
+
x = torch.concat([x1, x2], dim=2)
|
| 62 |
+
else:
|
| 63 |
+
x = x2
|
| 64 |
+
out = self.model(x).transpose(1, 2).contiguous()
|
| 65 |
+
return out, mel_len1 + mel_len2
|
cosyvoice/hifigan/f0_predictor.py
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Kai Hu)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
import torch
|
| 15 |
+
import torch.nn as nn
|
| 16 |
+
from torch.nn.utils import weight_norm
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
class ConvRNNF0Predictor(nn.Module):
|
| 20 |
+
def __init__(
|
| 21 |
+
self, num_class: int = 1, in_channels: int = 80, cond_channels: int = 512
|
| 22 |
+
):
|
| 23 |
+
super().__init__()
|
| 24 |
+
|
| 25 |
+
self.num_class = num_class
|
| 26 |
+
self.condnet = nn.Sequential(
|
| 27 |
+
weight_norm(
|
| 28 |
+
nn.Conv1d(in_channels, cond_channels, kernel_size=3, padding=1)
|
| 29 |
+
),
|
| 30 |
+
nn.ELU(),
|
| 31 |
+
weight_norm(
|
| 32 |
+
nn.Conv1d(cond_channels, cond_channels, kernel_size=3, padding=1)
|
| 33 |
+
),
|
| 34 |
+
nn.ELU(),
|
| 35 |
+
weight_norm(
|
| 36 |
+
nn.Conv1d(cond_channels, cond_channels, kernel_size=3, padding=1)
|
| 37 |
+
),
|
| 38 |
+
nn.ELU(),
|
| 39 |
+
weight_norm(
|
| 40 |
+
nn.Conv1d(cond_channels, cond_channels, kernel_size=3, padding=1)
|
| 41 |
+
),
|
| 42 |
+
nn.ELU(),
|
| 43 |
+
weight_norm(
|
| 44 |
+
nn.Conv1d(cond_channels, cond_channels, kernel_size=3, padding=1)
|
| 45 |
+
),
|
| 46 |
+
nn.ELU(),
|
| 47 |
+
)
|
| 48 |
+
self.classifier = nn.Linear(
|
| 49 |
+
in_features=cond_channels, out_features=self.num_class
|
| 50 |
+
)
|
| 51 |
+
|
| 52 |
+
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
| 53 |
+
x = self.condnet(x)
|
| 54 |
+
x = x.transpose(1, 2)
|
| 55 |
+
return torch.abs(self.classifier(x).squeeze(-1))
|
cosyvoice/hifigan/generator.py
ADDED
|
@@ -0,0 +1,566 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Kai Hu)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
|
| 15 |
+
"""HIFI-GAN"""
|
| 16 |
+
|
| 17 |
+
import typing as tp
|
| 18 |
+
import time
|
| 19 |
+
import numpy as np
|
| 20 |
+
from scipy.signal import get_window
|
| 21 |
+
import torch
|
| 22 |
+
import torch.nn as nn
|
| 23 |
+
import torch.nn.functional as F
|
| 24 |
+
from torch.nn import Conv1d
|
| 25 |
+
from torch.nn import ConvTranspose1d
|
| 26 |
+
from torch.nn.utils import remove_weight_norm
|
| 27 |
+
from torch.nn.utils import weight_norm
|
| 28 |
+
from torch.distributions.uniform import Uniform
|
| 29 |
+
|
| 30 |
+
from cosyvoice.transformer.activation import Snake
|
| 31 |
+
from cosyvoice.utils.common import get_padding
|
| 32 |
+
from cosyvoice.utils.common import init_weights
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
"""hifigan based generator implementation.
|
| 36 |
+
|
| 37 |
+
This code is modified from https://github.com/jik876/hifi-gan
|
| 38 |
+
,https://github.com/kan-bayashi/ParallelWaveGAN and
|
| 39 |
+
https://github.com/NVIDIA/BigVGAN
|
| 40 |
+
|
| 41 |
+
"""
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
class ResBlock(torch.nn.Module):
|
| 45 |
+
"""Residual block module in HiFiGAN/BigVGAN."""
|
| 46 |
+
|
| 47 |
+
def __init__(
|
| 48 |
+
self,
|
| 49 |
+
channels: int = 512,
|
| 50 |
+
kernel_size: int = 3,
|
| 51 |
+
dilations: tp.List[int] = [1, 3, 5],
|
| 52 |
+
):
|
| 53 |
+
super(ResBlock, self).__init__()
|
| 54 |
+
self.convs1 = nn.ModuleList()
|
| 55 |
+
self.convs2 = nn.ModuleList()
|
| 56 |
+
|
| 57 |
+
for dilation in dilations:
|
| 58 |
+
self.convs1.append(
|
| 59 |
+
weight_norm(
|
| 60 |
+
Conv1d(
|
| 61 |
+
channels,
|
| 62 |
+
channels,
|
| 63 |
+
kernel_size,
|
| 64 |
+
1,
|
| 65 |
+
dilation=dilation,
|
| 66 |
+
padding=get_padding(kernel_size, dilation),
|
| 67 |
+
)
|
| 68 |
+
)
|
| 69 |
+
)
|
| 70 |
+
self.convs2.append(
|
| 71 |
+
weight_norm(
|
| 72 |
+
Conv1d(
|
| 73 |
+
channels,
|
| 74 |
+
channels,
|
| 75 |
+
kernel_size,
|
| 76 |
+
1,
|
| 77 |
+
dilation=1,
|
| 78 |
+
padding=get_padding(kernel_size, 1),
|
| 79 |
+
)
|
| 80 |
+
)
|
| 81 |
+
)
|
| 82 |
+
self.convs1.apply(init_weights)
|
| 83 |
+
self.convs2.apply(init_weights)
|
| 84 |
+
self.activations1 = nn.ModuleList(
|
| 85 |
+
[Snake(channels, alpha_logscale=False) for _ in range(len(self.convs1))]
|
| 86 |
+
)
|
| 87 |
+
self.activations2 = nn.ModuleList(
|
| 88 |
+
[Snake(channels, alpha_logscale=False) for _ in range(len(self.convs2))]
|
| 89 |
+
)
|
| 90 |
+
|
| 91 |
+
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
| 92 |
+
for idx in range(len(self.convs1)):
|
| 93 |
+
xt = self.activations1[idx](x)
|
| 94 |
+
xt = self.convs1[idx](xt)
|
| 95 |
+
xt = self.activations2[idx](xt)
|
| 96 |
+
xt = self.convs2[idx](xt)
|
| 97 |
+
x = xt + x
|
| 98 |
+
return x
|
| 99 |
+
|
| 100 |
+
def remove_weight_norm(self):
|
| 101 |
+
for idx in range(len(self.convs1)):
|
| 102 |
+
remove_weight_norm(self.convs1[idx])
|
| 103 |
+
remove_weight_norm(self.convs2[idx])
|
| 104 |
+
|
| 105 |
+
|
| 106 |
+
class SineGen(torch.nn.Module):
|
| 107 |
+
"""Definition of sine generator
|
| 108 |
+
SineGen(samp_rate, harmonic_num = 0,
|
| 109 |
+
sine_amp = 0.1, noise_std = 0.003,
|
| 110 |
+
voiced_threshold = 0,
|
| 111 |
+
flag_for_pulse=False)
|
| 112 |
+
samp_rate: sampling rate in Hz
|
| 113 |
+
harmonic_num: number of harmonic overtones (default 0)
|
| 114 |
+
sine_amp: amplitude of sine-wavefrom (default 0.1)
|
| 115 |
+
noise_std: std of Gaussian noise (default 0.003)
|
| 116 |
+
voiced_thoreshold: F0 threshold for U/V classification (default 0)
|
| 117 |
+
flag_for_pulse: this SinGen is used inside PulseGen (default False)
|
| 118 |
+
Note: when flag_for_pulse is True, the first time step of a voiced
|
| 119 |
+
segment is always sin(np.pi) or cos(0)
|
| 120 |
+
"""
|
| 121 |
+
|
| 122 |
+
def __init__(
|
| 123 |
+
self,
|
| 124 |
+
samp_rate,
|
| 125 |
+
harmonic_num=0,
|
| 126 |
+
sine_amp=0.1,
|
| 127 |
+
noise_std=0.003,
|
| 128 |
+
voiced_threshold=0,
|
| 129 |
+
):
|
| 130 |
+
super(SineGen, self).__init__()
|
| 131 |
+
self.sine_amp = sine_amp
|
| 132 |
+
self.noise_std = noise_std
|
| 133 |
+
self.harmonic_num = harmonic_num
|
| 134 |
+
self.sampling_rate = samp_rate
|
| 135 |
+
self.voiced_threshold = voiced_threshold
|
| 136 |
+
|
| 137 |
+
def _f02uv(self, f0):
|
| 138 |
+
# generate uv signal
|
| 139 |
+
uv = (f0 > self.voiced_threshold).type(torch.float32)
|
| 140 |
+
return uv
|
| 141 |
+
|
| 142 |
+
@torch.no_grad()
|
| 143 |
+
def forward(self, f0):
|
| 144 |
+
"""
|
| 145 |
+
:param f0: [B, 1, sample_len], Hz
|
| 146 |
+
:return: [B, 1, sample_len]
|
| 147 |
+
"""
|
| 148 |
+
|
| 149 |
+
F_mat = torch.zeros((f0.size(0), self.harmonic_num + 1, f0.size(-1))).to(
|
| 150 |
+
f0.device
|
| 151 |
+
)
|
| 152 |
+
for i in range(self.harmonic_num + 1):
|
| 153 |
+
F_mat[:, i : i + 1, :] = f0 * (i + 1) / self.sampling_rate
|
| 154 |
+
|
| 155 |
+
theta_mat = 2 * np.pi * (torch.cumsum(F_mat, dim=-1) % 1)
|
| 156 |
+
u_dist = Uniform(low=-np.pi, high=np.pi)
|
| 157 |
+
phase_vec = u_dist.sample(
|
| 158 |
+
sample_shape=(f0.size(0), self.harmonic_num + 1, 1)
|
| 159 |
+
).to(F_mat.device)
|
| 160 |
+
phase_vec[:, 0, :] = 0
|
| 161 |
+
|
| 162 |
+
# generate sine waveforms
|
| 163 |
+
sine_waves = self.sine_amp * torch.sin(theta_mat + phase_vec)
|
| 164 |
+
|
| 165 |
+
# generate uv signal
|
| 166 |
+
uv = self._f02uv(f0)
|
| 167 |
+
|
| 168 |
+
# noise: for unvoiced should be similar to sine_amp
|
| 169 |
+
# std = self.sine_amp/3 -> max value ~ self.sine_amp
|
| 170 |
+
# . for voiced regions is self.noise_std
|
| 171 |
+
noise_amp = uv * self.noise_std + (1 - uv) * self.sine_amp / 3
|
| 172 |
+
noise = noise_amp * torch.randn_like(sine_waves)
|
| 173 |
+
|
| 174 |
+
# first: set the unvoiced part to 0 by uv
|
| 175 |
+
# then: additive noise
|
| 176 |
+
sine_waves = sine_waves * uv + noise
|
| 177 |
+
return sine_waves, uv, noise
|
| 178 |
+
|
| 179 |
+
|
| 180 |
+
class SourceModuleHnNSF(torch.nn.Module):
|
| 181 |
+
"""SourceModule for hn-nsf
|
| 182 |
+
SourceModule(sampling_rate, harmonic_num=0, sine_amp=0.1,
|
| 183 |
+
add_noise_std=0.003, voiced_threshod=0)
|
| 184 |
+
sampling_rate: sampling_rate in Hz
|
| 185 |
+
harmonic_num: number of harmonic above F0 (default: 0)
|
| 186 |
+
sine_amp: amplitude of sine source signal (default: 0.1)
|
| 187 |
+
add_noise_std: std of additive Gaussian noise (default: 0.003)
|
| 188 |
+
note that amplitude of noise in unvoiced is decided
|
| 189 |
+
by sine_amp
|
| 190 |
+
voiced_threshold: threhold to set U/V given F0 (default: 0)
|
| 191 |
+
Sine_source, noise_source = SourceModuleHnNSF(F0_sampled)
|
| 192 |
+
F0_sampled (batchsize, length, 1)
|
| 193 |
+
Sine_source (batchsize, length, 1)
|
| 194 |
+
noise_source (batchsize, length 1)
|
| 195 |
+
uv (batchsize, length, 1)
|
| 196 |
+
"""
|
| 197 |
+
|
| 198 |
+
def __init__(
|
| 199 |
+
self,
|
| 200 |
+
sampling_rate,
|
| 201 |
+
upsample_scale,
|
| 202 |
+
harmonic_num=0,
|
| 203 |
+
sine_amp=0.1,
|
| 204 |
+
add_noise_std=0.003,
|
| 205 |
+
voiced_threshod=0,
|
| 206 |
+
):
|
| 207 |
+
super(SourceModuleHnNSF, self).__init__()
|
| 208 |
+
|
| 209 |
+
self.sine_amp = sine_amp
|
| 210 |
+
self.noise_std = add_noise_std
|
| 211 |
+
|
| 212 |
+
# to produce sine waveforms
|
| 213 |
+
self.l_sin_gen = SineGen(
|
| 214 |
+
sampling_rate, harmonic_num, sine_amp, add_noise_std, voiced_threshod
|
| 215 |
+
)
|
| 216 |
+
|
| 217 |
+
# to merge source harmonics into a single excitation
|
| 218 |
+
self.l_linear = torch.nn.Linear(harmonic_num + 1, 1)
|
| 219 |
+
self.l_tanh = torch.nn.Tanh()
|
| 220 |
+
|
| 221 |
+
def forward(self, x):
|
| 222 |
+
"""
|
| 223 |
+
Sine_source, noise_source = SourceModuleHnNSF(F0_sampled)
|
| 224 |
+
F0_sampled (batchsize, length, 1)
|
| 225 |
+
Sine_source (batchsize, length, 1)
|
| 226 |
+
noise_source (batchsize, length 1)
|
| 227 |
+
"""
|
| 228 |
+
# source for harmonic branch
|
| 229 |
+
with torch.no_grad():
|
| 230 |
+
sine_wavs, uv, _ = self.l_sin_gen(x.transpose(1, 2))
|
| 231 |
+
sine_wavs = sine_wavs.transpose(1, 2)
|
| 232 |
+
uv = uv.transpose(1, 2)
|
| 233 |
+
sine_merge = self.l_tanh(self.l_linear(sine_wavs))
|
| 234 |
+
|
| 235 |
+
# source for noise branch, in the same shape as uv
|
| 236 |
+
noise = torch.randn_like(uv) * self.sine_amp / 3
|
| 237 |
+
return sine_merge, noise, uv
|
| 238 |
+
|
| 239 |
+
|
| 240 |
+
class HiFTGenerator(nn.Module):
|
| 241 |
+
"""
|
| 242 |
+
HiFTNet Generator: Neural Source Filter + ISTFTNet
|
| 243 |
+
https://arxiv.org/abs/2309.09493
|
| 244 |
+
"""
|
| 245 |
+
|
| 246 |
+
def __init__(
|
| 247 |
+
self,
|
| 248 |
+
in_channels: int = 80,
|
| 249 |
+
base_channels: int = 512,
|
| 250 |
+
nb_harmonics: int = 8,
|
| 251 |
+
sampling_rate: int = 22050,
|
| 252 |
+
nsf_alpha: float = 0.1,
|
| 253 |
+
nsf_sigma: float = 0.003,
|
| 254 |
+
nsf_voiced_threshold: float = 10,
|
| 255 |
+
upsample_rates: tp.List[int] = [8, 8],
|
| 256 |
+
upsample_kernel_sizes: tp.List[int] = [16, 16],
|
| 257 |
+
istft_params: tp.Dict[str, int] = {"n_fft": 16, "hop_len": 4},
|
| 258 |
+
resblock_kernel_sizes: tp.List[int] = [3, 7, 11],
|
| 259 |
+
resblock_dilation_sizes: tp.List[tp.List[int]] = [
|
| 260 |
+
[1, 3, 5],
|
| 261 |
+
[1, 3, 5],
|
| 262 |
+
[1, 3, 5],
|
| 263 |
+
],
|
| 264 |
+
source_resblock_kernel_sizes: tp.List[int] = [7, 11],
|
| 265 |
+
source_resblock_dilation_sizes: tp.List[tp.List[int]] = [[1, 3, 5], [1, 3, 5]],
|
| 266 |
+
lrelu_slope: float = 0.1,
|
| 267 |
+
audio_limit: float = 0.99,
|
| 268 |
+
f0_predictor: torch.nn.Module = None,
|
| 269 |
+
):
|
| 270 |
+
super(HiFTGenerator, self).__init__()
|
| 271 |
+
|
| 272 |
+
self.out_channels = 1
|
| 273 |
+
self.nb_harmonics = nb_harmonics
|
| 274 |
+
self.sampling_rate = sampling_rate
|
| 275 |
+
self.istft_params = istft_params
|
| 276 |
+
self.lrelu_slope = lrelu_slope
|
| 277 |
+
self.audio_limit = audio_limit
|
| 278 |
+
|
| 279 |
+
self.num_kernels = len(resblock_kernel_sizes)
|
| 280 |
+
self.num_upsamples = len(upsample_rates)
|
| 281 |
+
self.upsample_rates = upsample_rates
|
| 282 |
+
self.m_source = SourceModuleHnNSF(
|
| 283 |
+
sampling_rate=sampling_rate,
|
| 284 |
+
upsample_scale=np.prod(upsample_rates) * istft_params["hop_len"],
|
| 285 |
+
harmonic_num=nb_harmonics,
|
| 286 |
+
sine_amp=nsf_alpha,
|
| 287 |
+
add_noise_std=nsf_sigma,
|
| 288 |
+
voiced_threshod=nsf_voiced_threshold,
|
| 289 |
+
)
|
| 290 |
+
self.f0_upsamp = torch.nn.Upsample(
|
| 291 |
+
scale_factor=np.prod(upsample_rates) * istft_params["hop_len"]
|
| 292 |
+
)
|
| 293 |
+
|
| 294 |
+
self.conv_pre = weight_norm(Conv1d(in_channels, base_channels, 7, 1, padding=3))
|
| 295 |
+
|
| 296 |
+
# Up
|
| 297 |
+
self.ups = nn.ModuleList()
|
| 298 |
+
for i, (u, k) in enumerate(zip(upsample_rates, upsample_kernel_sizes)):
|
| 299 |
+
self.ups.append(
|
| 300 |
+
weight_norm(
|
| 301 |
+
ConvTranspose1d(
|
| 302 |
+
base_channels // (2**i),
|
| 303 |
+
base_channels // (2 ** (i + 1)),
|
| 304 |
+
k,
|
| 305 |
+
u,
|
| 306 |
+
padding=(k - u) // 2,
|
| 307 |
+
)
|
| 308 |
+
)
|
| 309 |
+
)
|
| 310 |
+
|
| 311 |
+
# Down
|
| 312 |
+
self.source_downs = nn.ModuleList()
|
| 313 |
+
self.source_resblocks = nn.ModuleList()
|
| 314 |
+
downsample_rates = [1] + upsample_rates[::-1][:-1]
|
| 315 |
+
downsample_cum_rates = np.cumprod(downsample_rates)
|
| 316 |
+
for i, (u, k, d) in enumerate(
|
| 317 |
+
zip(
|
| 318 |
+
downsample_cum_rates[::-1],
|
| 319 |
+
source_resblock_kernel_sizes,
|
| 320 |
+
source_resblock_dilation_sizes,
|
| 321 |
+
)
|
| 322 |
+
):
|
| 323 |
+
if u == 1:
|
| 324 |
+
self.source_downs.append(
|
| 325 |
+
Conv1d(
|
| 326 |
+
istft_params["n_fft"] + 2, base_channels // (2 ** (i + 1)), 1, 1
|
| 327 |
+
)
|
| 328 |
+
)
|
| 329 |
+
else:
|
| 330 |
+
self.source_downs.append(
|
| 331 |
+
Conv1d(
|
| 332 |
+
istft_params["n_fft"] + 2,
|
| 333 |
+
base_channels // (2 ** (i + 1)),
|
| 334 |
+
u * 2,
|
| 335 |
+
u,
|
| 336 |
+
padding=(u // 2),
|
| 337 |
+
)
|
| 338 |
+
)
|
| 339 |
+
|
| 340 |
+
self.source_resblocks.append(
|
| 341 |
+
ResBlock(base_channels // (2 ** (i + 1)), k, d)
|
| 342 |
+
)
|
| 343 |
+
|
| 344 |
+
self.resblocks = nn.ModuleList()
|
| 345 |
+
for i in range(len(self.ups)):
|
| 346 |
+
ch = base_channels // (2 ** (i + 1))
|
| 347 |
+
for _, (k, d) in enumerate(
|
| 348 |
+
zip(resblock_kernel_sizes, resblock_dilation_sizes)
|
| 349 |
+
):
|
| 350 |
+
self.resblocks.append(ResBlock(ch, k, d))
|
| 351 |
+
|
| 352 |
+
self.conv_post = weight_norm(
|
| 353 |
+
Conv1d(ch, istft_params["n_fft"] + 2, 7, 1, padding=3)
|
| 354 |
+
)
|
| 355 |
+
self.ups.apply(init_weights)
|
| 356 |
+
self.conv_post.apply(init_weights)
|
| 357 |
+
self.reflection_pad = nn.ReflectionPad1d((1, 0))
|
| 358 |
+
self.stft_window = torch.from_numpy(
|
| 359 |
+
get_window("hann", istft_params["n_fft"], fftbins=True).astype(np.float32)
|
| 360 |
+
).cuda()
|
| 361 |
+
self.f0_predictor = f0_predictor
|
| 362 |
+
self.inference_buffers = {}
|
| 363 |
+
self.inference_graphs = {}
|
| 364 |
+
|
| 365 |
+
def _f02source(self, f0: torch.Tensor) -> torch.Tensor:
|
| 366 |
+
f0 = self.f0_upsamp(f0[:, None]).transpose(1, 2) # bs,n,t
|
| 367 |
+
|
| 368 |
+
har_source, _, _ = self.m_source(f0)
|
| 369 |
+
return har_source.transpose(1, 2)
|
| 370 |
+
|
| 371 |
+
def _stft(self, x):
|
| 372 |
+
spec = torch.stft(
|
| 373 |
+
x,
|
| 374 |
+
self.istft_params["n_fft"],
|
| 375 |
+
self.istft_params["hop_len"],
|
| 376 |
+
self.istft_params["n_fft"],
|
| 377 |
+
window=self.stft_window,
|
| 378 |
+
return_complex=True,
|
| 379 |
+
)
|
| 380 |
+
spec = torch.view_as_real(spec) # [B, F, TT, 2]
|
| 381 |
+
return spec[..., 0], spec[..., 1]
|
| 382 |
+
|
| 383 |
+
def _istft(self, magnitude, phase):
|
| 384 |
+
magnitude = torch.clip(magnitude, max=1e2)
|
| 385 |
+
real = magnitude * torch.cos(phase)
|
| 386 |
+
img = magnitude * torch.sin(phase)
|
| 387 |
+
inverse_transform = torch.istft(
|
| 388 |
+
torch.complex(real, img),
|
| 389 |
+
self.istft_params["n_fft"],
|
| 390 |
+
self.istft_params["hop_len"],
|
| 391 |
+
self.istft_params["n_fft"],
|
| 392 |
+
window=self.stft_window,
|
| 393 |
+
)
|
| 394 |
+
return inverse_transform
|
| 395 |
+
|
| 396 |
+
def forward(
|
| 397 |
+
self, x: torch.Tensor, cache_source: torch.Tensor = torch.zeros(1, 1, 0)
|
| 398 |
+
) -> torch.Tensor:
|
| 399 |
+
f0 = self.f0_predictor(x)
|
| 400 |
+
s = self._f02source(f0)
|
| 401 |
+
|
| 402 |
+
# use cache_source to avoid glitch
|
| 403 |
+
if cache_source.shape[2] != 0:
|
| 404 |
+
s[:, :, : cache_source.shape[2]] = cache_source
|
| 405 |
+
|
| 406 |
+
s_stft_real, s_stft_imag = self._stft(s.squeeze(1))
|
| 407 |
+
s_stft = torch.cat([s_stft_real, s_stft_imag], dim=1)
|
| 408 |
+
|
| 409 |
+
x = self.conv_pre(x)
|
| 410 |
+
for i in range(self.num_upsamples):
|
| 411 |
+
x = F.leaky_relu(x, self.lrelu_slope)
|
| 412 |
+
x = self.ups[i](x)
|
| 413 |
+
|
| 414 |
+
if i == self.num_upsamples - 1:
|
| 415 |
+
x = self.reflection_pad(x)
|
| 416 |
+
|
| 417 |
+
# fusion
|
| 418 |
+
si = self.source_downs[i](s_stft)
|
| 419 |
+
si = self.source_resblocks[i](si)
|
| 420 |
+
x = x + si
|
| 421 |
+
|
| 422 |
+
xs = None
|
| 423 |
+
for j in range(self.num_kernels):
|
| 424 |
+
if xs is None:
|
| 425 |
+
xs = self.resblocks[i * self.num_kernels + j](x)
|
| 426 |
+
else:
|
| 427 |
+
xs += self.resblocks[i * self.num_kernels + j](x)
|
| 428 |
+
x = xs / self.num_kernels
|
| 429 |
+
|
| 430 |
+
x = F.leaky_relu(x)
|
| 431 |
+
x = self.conv_post(x)
|
| 432 |
+
magnitude = torch.exp(x[:, : self.istft_params["n_fft"] // 2 + 1, :])
|
| 433 |
+
phase = torch.sin(
|
| 434 |
+
x[:, self.istft_params["n_fft"] // 2 + 1 :, :]
|
| 435 |
+
) # actually, sin is redundancy
|
| 436 |
+
|
| 437 |
+
x = self._istft(magnitude, phase)
|
| 438 |
+
x = torch.clamp(x, -self.audio_limit, self.audio_limit)
|
| 439 |
+
return x, s
|
| 440 |
+
|
| 441 |
+
def remove_weight_norm(self):
|
| 442 |
+
print("Removing weight norm...")
|
| 443 |
+
for l in self.ups:
|
| 444 |
+
remove_weight_norm(l)
|
| 445 |
+
for l in self.resblocks:
|
| 446 |
+
l.remove_weight_norm()
|
| 447 |
+
remove_weight_norm(self.conv_pre)
|
| 448 |
+
remove_weight_norm(self.conv_post)
|
| 449 |
+
self.source_module.remove_weight_norm()
|
| 450 |
+
for l in self.source_downs:
|
| 451 |
+
remove_weight_norm(l)
|
| 452 |
+
for l in self.source_resblocks:
|
| 453 |
+
l.remove_weight_norm()
|
| 454 |
+
|
| 455 |
+
@torch.inference_mode()
|
| 456 |
+
def _inference_impl(self, mel: torch.Tensor, s_stft: torch.Tensor) -> torch.Tensor:
|
| 457 |
+
x = self.conv_pre(mel)
|
| 458 |
+
for i in range(self.num_upsamples):
|
| 459 |
+
x = F.leaky_relu(x, self.lrelu_slope)
|
| 460 |
+
x = self.ups[i](x)
|
| 461 |
+
|
| 462 |
+
if i == self.num_upsamples - 1:
|
| 463 |
+
x = self.reflection_pad(x)
|
| 464 |
+
|
| 465 |
+
# fusion
|
| 466 |
+
si = self.source_downs[i](s_stft)
|
| 467 |
+
si = self.source_resblocks[i](si)
|
| 468 |
+
x = x + si
|
| 469 |
+
|
| 470 |
+
xs = None
|
| 471 |
+
for j in range(self.num_kernels):
|
| 472 |
+
if xs is None:
|
| 473 |
+
xs = self.resblocks[i * self.num_kernels + j](x)
|
| 474 |
+
else:
|
| 475 |
+
xs += self.resblocks[i * self.num_kernels + j](x)
|
| 476 |
+
x = xs / self.num_kernels
|
| 477 |
+
|
| 478 |
+
x = F.leaky_relu(x)
|
| 479 |
+
x = self.conv_post(x)
|
| 480 |
+
magnitude = torch.exp(x[:, : self.istft_params["n_fft"] // 2 + 1, :])
|
| 481 |
+
phase = torch.sin(
|
| 482 |
+
x[:, self.istft_params["n_fft"] // 2 + 1 :, :]
|
| 483 |
+
) # actually, sin is redundancy
|
| 484 |
+
# print(f"mel: {mel.shape}, magnitude: {magnitude.shape}, phase: {phase.shape}")
|
| 485 |
+
return magnitude, phase
|
| 486 |
+
|
| 487 |
+
@torch.inference_mode()
|
| 488 |
+
def inference(
|
| 489 |
+
self, mel: torch.Tensor, cache_source: torch.Tensor = torch.zeros(1, 1, 0)
|
| 490 |
+
) -> torch.Tensor:
|
| 491 |
+
curr_seq_len = mel.shape[2]
|
| 492 |
+
f0 = self.f0_predictor(mel)
|
| 493 |
+
s = self._f02source(f0)
|
| 494 |
+
s_stft_real, s_stft_imag = self._stft(s.squeeze(1))
|
| 495 |
+
s_stft = torch.cat([s_stft_real, s_stft_imag], dim=1)
|
| 496 |
+
|
| 497 |
+
target_len = None
|
| 498 |
+
for seq_len in sorted(self.inference_buffers.keys()):
|
| 499 |
+
if curr_seq_len <= seq_len:
|
| 500 |
+
target_len = seq_len
|
| 501 |
+
break
|
| 502 |
+
|
| 503 |
+
if target_len is not None:
|
| 504 |
+
buffer = self.inference_buffers[target_len]
|
| 505 |
+
|
| 506 |
+
if curr_seq_len < target_len:
|
| 507 |
+
padded_mel = torch.zeros_like(buffer["mel"])
|
| 508 |
+
padded_mel[:, :, :curr_seq_len] = mel
|
| 509 |
+
buffer["mel"].copy_(padded_mel)
|
| 510 |
+
padded_s_stft = torch.zeros_like(buffer["s_stft"])
|
| 511 |
+
cur_s_stft_len = s_stft.shape[2]
|
| 512 |
+
padded_s_stft[:, :, :cur_s_stft_len] = s_stft
|
| 513 |
+
buffer["s_stft"].copy_(padded_s_stft)
|
| 514 |
+
|
| 515 |
+
else:
|
| 516 |
+
buffer["mel"].copy_(mel)
|
| 517 |
+
buffer["s_stft"].copy_(s_stft)
|
| 518 |
+
cur_s_stft_len = s_stft.shape[2]
|
| 519 |
+
|
| 520 |
+
self.inference_graphs[target_len].replay()
|
| 521 |
+
|
| 522 |
+
magnitude, phase = (
|
| 523 |
+
buffer["magnitude"][:, :, :cur_s_stft_len],
|
| 524 |
+
buffer["phase"][:, :, :cur_s_stft_len],
|
| 525 |
+
)
|
| 526 |
+
else:
|
| 527 |
+
magnitude, phase = self._inference_impl(mel=mel, s_stft=s_stft)
|
| 528 |
+
|
| 529 |
+
x = self._istft(magnitude, phase)
|
| 530 |
+
x = torch.clamp(x, -self.audio_limit, self.audio_limit)
|
| 531 |
+
return x, s
|
| 532 |
+
|
| 533 |
+
@torch.inference_mode()
|
| 534 |
+
def capture_inference(self, seq_len_to_capture=[64, 128, 256, 512, 1024]):
|
| 535 |
+
start_time = time.time()
|
| 536 |
+
print(
|
| 537 |
+
f"capture inference for HiFTGenerator with seq_len_to_capture: {seq_len_to_capture}"
|
| 538 |
+
)
|
| 539 |
+
for seq_len in seq_len_to_capture:
|
| 540 |
+
mel = torch.randn(
|
| 541 |
+
1, 80, seq_len, device=torch.device("cuda"), dtype=torch.float32
|
| 542 |
+
)
|
| 543 |
+
f0 = self.f0_predictor(mel)
|
| 544 |
+
s = self._f02source(f0)
|
| 545 |
+
s_stft_real, s_stft_imag = self._stft(s.squeeze(1))
|
| 546 |
+
s_stft = torch.cat([s_stft_real, s_stft_imag], dim=1)
|
| 547 |
+
|
| 548 |
+
magnitude, phase = self._inference_impl(mel=mel, s_stft=s_stft)
|
| 549 |
+
torch.cuda.synchronize()
|
| 550 |
+
|
| 551 |
+
g = torch.cuda.CUDAGraph()
|
| 552 |
+
with torch.cuda.graph(g):
|
| 553 |
+
magnitude, phase = self._inference_impl(mel=mel, s_stft=s_stft)
|
| 554 |
+
inference_buffer = {
|
| 555 |
+
"mel": mel,
|
| 556 |
+
"s_stft": s_stft,
|
| 557 |
+
"magnitude": magnitude,
|
| 558 |
+
"phase": phase,
|
| 559 |
+
}
|
| 560 |
+
self.inference_buffers[seq_len] = inference_buffer
|
| 561 |
+
self.inference_graphs[seq_len] = g
|
| 562 |
+
|
| 563 |
+
end_time = time.time()
|
| 564 |
+
print(
|
| 565 |
+
f"capture inference for HiFTGenerator with seq_len_to_capture: {seq_len_to_capture} takes {end_time - start_time} seconds"
|
| 566 |
+
)
|
cosyvoice/matcha/audio.py
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import numpy as np
|
| 2 |
+
import torch
|
| 3 |
+
import torch.utils.data
|
| 4 |
+
from librosa.filters import mel as librosa_mel_fn
|
| 5 |
+
from scipy.io.wavfile import read
|
| 6 |
+
|
| 7 |
+
MAX_WAV_VALUE = 32768.0
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def load_wav(full_path):
|
| 11 |
+
sampling_rate, data = read(full_path)
|
| 12 |
+
return data, sampling_rate
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def dynamic_range_compression(x, C=1, clip_val=1e-5):
|
| 16 |
+
return np.log(np.clip(x, a_min=clip_val, a_max=None) * C)
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
def dynamic_range_decompression(x, C=1):
|
| 20 |
+
return np.exp(x) / C
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
def dynamic_range_compression_torch(x, C=1, clip_val=1e-5):
|
| 24 |
+
return torch.log(torch.clamp(x, min=clip_val) * C)
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
def dynamic_range_decompression_torch(x, C=1):
|
| 28 |
+
return torch.exp(x) / C
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
def spectral_normalize_torch(magnitudes):
|
| 32 |
+
output = dynamic_range_compression_torch(magnitudes)
|
| 33 |
+
return output
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
def spectral_de_normalize_torch(magnitudes):
|
| 37 |
+
output = dynamic_range_decompression_torch(magnitudes)
|
| 38 |
+
return output
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
mel_basis = {}
|
| 42 |
+
hann_window = {}
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
def mel_spectrogram(
|
| 46 |
+
y, n_fft, num_mels, sampling_rate, hop_size, win_size, fmin, fmax, center=False
|
| 47 |
+
):
|
| 48 |
+
if torch.min(y) < -1.0:
|
| 49 |
+
print("min value is ", torch.min(y))
|
| 50 |
+
if torch.max(y) > 1.0:
|
| 51 |
+
print("max value is ", torch.max(y))
|
| 52 |
+
|
| 53 |
+
global mel_basis, hann_window # pylint: disable=global-statement
|
| 54 |
+
if f"{str(fmax)}_{str(y.device)}" not in mel_basis:
|
| 55 |
+
mel = librosa_mel_fn(
|
| 56 |
+
sr=sampling_rate, n_fft=n_fft, n_mels=num_mels, fmin=fmin, fmax=fmax
|
| 57 |
+
)
|
| 58 |
+
mel_basis[str(fmax) + "_" + str(y.device)] = (
|
| 59 |
+
torch.from_numpy(mel).float().to(y.device)
|
| 60 |
+
)
|
| 61 |
+
hann_window[str(y.device)] = torch.hann_window(win_size).to(y.device)
|
| 62 |
+
|
| 63 |
+
y = torch.nn.functional.pad(
|
| 64 |
+
y.unsqueeze(1),
|
| 65 |
+
(int((n_fft - hop_size) / 2), int((n_fft - hop_size) / 2)),
|
| 66 |
+
mode="reflect",
|
| 67 |
+
)
|
| 68 |
+
y = y.squeeze(1)
|
| 69 |
+
|
| 70 |
+
spec = torch.view_as_real(
|
| 71 |
+
torch.stft(
|
| 72 |
+
y,
|
| 73 |
+
n_fft,
|
| 74 |
+
hop_length=hop_size,
|
| 75 |
+
win_length=win_size,
|
| 76 |
+
window=hann_window[str(y.device)],
|
| 77 |
+
center=center,
|
| 78 |
+
pad_mode="reflect",
|
| 79 |
+
normalized=False,
|
| 80 |
+
onesided=True,
|
| 81 |
+
return_complex=True,
|
| 82 |
+
)
|
| 83 |
+
)
|
| 84 |
+
|
| 85 |
+
spec = torch.sqrt(spec.pow(2).sum(-1) + (1e-9))
|
| 86 |
+
|
| 87 |
+
spec = torch.matmul(mel_basis[str(fmax) + "_" + str(y.device)], spec)
|
| 88 |
+
spec = spectral_normalize_torch(spec)
|
| 89 |
+
|
| 90 |
+
return spec
|
cosyvoice/matcha/decoder.py
ADDED
|
@@ -0,0 +1,511 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import math
|
| 2 |
+
from typing import Optional
|
| 3 |
+
|
| 4 |
+
import torch
|
| 5 |
+
import torch.nn as nn
|
| 6 |
+
import torch.nn.functional as F
|
| 7 |
+
from conformer import ConformerBlock
|
| 8 |
+
from diffusers.models.activations import get_activation
|
| 9 |
+
from einops import pack, rearrange, repeat
|
| 10 |
+
|
| 11 |
+
from cosyvoice.matcha.transformer import BasicTransformerBlock
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
class SinusoidalPosEmb(torch.nn.Module):
|
| 15 |
+
def __init__(self, dim):
|
| 16 |
+
super().__init__()
|
| 17 |
+
self.dim = dim
|
| 18 |
+
assert self.dim % 2 == 0, "SinusoidalPosEmb requires dim to be even"
|
| 19 |
+
|
| 20 |
+
def forward(self, x, scale=1000):
|
| 21 |
+
if x.ndim < 1:
|
| 22 |
+
x = x.unsqueeze(0)
|
| 23 |
+
device = x.device
|
| 24 |
+
half_dim = self.dim // 2
|
| 25 |
+
emb = math.log(10000) / (half_dim - 1)
|
| 26 |
+
emb = torch.exp(torch.arange(half_dim, device=device).float() * -emb)
|
| 27 |
+
emb = scale * x.unsqueeze(1) * emb.unsqueeze(0)
|
| 28 |
+
emb = torch.cat((emb.sin(), emb.cos()), dim=-1)
|
| 29 |
+
return emb
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
class MaskedGroupNorm(nn.GroupNorm):
|
| 33 |
+
"""
|
| 34 |
+
Masked verstion of the Group normalization.
|
| 35 |
+
|
| 36 |
+
Based on: https://github.com/ptrblck/pytorch_misc/blob/20e8ea93bd458b88f921a87e2d4001a4eb753a02/batch_norm_manual.py
|
| 37 |
+
|
| 38 |
+
Receives a N-dim tensor of sequence lengths per batch element
|
| 39 |
+
along with the regular input for masking.
|
| 40 |
+
|
| 41 |
+
Check pytorch's GroupNorm implementation for argument details.
|
| 42 |
+
"""
|
| 43 |
+
|
| 44 |
+
def __init__(self, num_groups, num_channels, eps=1e-5, affine=True):
|
| 45 |
+
super(MaskedGroupNorm, self).__init__(num_groups, num_channels, eps, affine)
|
| 46 |
+
|
| 47 |
+
def forward(self, inp, mask=None):
|
| 48 |
+
assert (
|
| 49 |
+
inp.shape[1] % self.num_groups == 0
|
| 50 |
+
), "Feature size not divisible by groups"
|
| 51 |
+
|
| 52 |
+
# 计算有效长度
|
| 53 |
+
seq_lengths = mask.sum(-1, keepdim=True) # [batch_size, 1]
|
| 54 |
+
|
| 55 |
+
# 将输入reshape为groups
|
| 56 |
+
features_per_group = inp.shape[1] // self.num_groups
|
| 57 |
+
inp_r = inp.reshape(
|
| 58 |
+
inp.shape[0], self.num_groups, features_per_group, inp.shape[-1]
|
| 59 |
+
)
|
| 60 |
+
mask_r = mask.unsqueeze(1) # [batch_size, 1, 1, length]
|
| 61 |
+
|
| 62 |
+
# 计算masked mean和variance
|
| 63 |
+
masked_inp = inp_r * mask_r
|
| 64 |
+
n = seq_lengths * features_per_group # 每组的有效元素数量
|
| 65 |
+
mean = masked_inp.sum([2, 3], keepdim=True) / (n.view(-1, 1, 1, 1) + 1e-5)
|
| 66 |
+
var = ((masked_inp - mean * mask_r) ** 2).sum([2, 3], keepdim=True) / (
|
| 67 |
+
n.view(-1, 1, 1, 1) + 1e-5
|
| 68 |
+
)
|
| 69 |
+
|
| 70 |
+
# 标准化
|
| 71 |
+
inp_r = (inp_r - mean) / (torch.sqrt(var + self.eps))
|
| 72 |
+
out = inp_r.reshape(inp.shape[0], self.num_channels, inp.shape[-1])
|
| 73 |
+
|
| 74 |
+
# 应用仿射变换
|
| 75 |
+
if self.affine:
|
| 76 |
+
out = out * self.weight[None, :, None] + self.bias[None, :, None]
|
| 77 |
+
|
| 78 |
+
return out
|
| 79 |
+
|
| 80 |
+
|
| 81 |
+
class Block1D(torch.nn.Module):
|
| 82 |
+
def __init__(self, dim, dim_out, groups=8):
|
| 83 |
+
super().__init__()
|
| 84 |
+
self.block = torch.nn.Sequential(
|
| 85 |
+
torch.nn.Conv1d(dim, dim_out, 3, padding=1),
|
| 86 |
+
torch.nn.GroupNorm(groups, dim_out),
|
| 87 |
+
# MaskedGroupNorm(groups, dim_out),
|
| 88 |
+
nn.Mish(),
|
| 89 |
+
)
|
| 90 |
+
|
| 91 |
+
def forward(self, x, mask):
|
| 92 |
+
output = self.block(x * mask)
|
| 93 |
+
return output * mask
|
| 94 |
+
return x * mask
|
| 95 |
+
|
| 96 |
+
|
| 97 |
+
class ResnetBlock1D(torch.nn.Module):
|
| 98 |
+
def __init__(self, dim, dim_out, time_emb_dim, groups=8):
|
| 99 |
+
super().__init__()
|
| 100 |
+
self.mlp = torch.nn.Sequential(
|
| 101 |
+
nn.Mish(), torch.nn.Linear(time_emb_dim, dim_out)
|
| 102 |
+
)
|
| 103 |
+
|
| 104 |
+
self.block1 = Block1D(dim, dim_out, groups=groups)
|
| 105 |
+
self.block2 = Block1D(dim_out, dim_out, groups=groups)
|
| 106 |
+
|
| 107 |
+
self.res_conv = torch.nn.Conv1d(dim, dim_out, 1)
|
| 108 |
+
|
| 109 |
+
def forward(self, x, mask, time_emb):
|
| 110 |
+
h = self.block1(x, mask)
|
| 111 |
+
h += self.mlp(time_emb).unsqueeze(-1)
|
| 112 |
+
h = self.block2(h, mask)
|
| 113 |
+
output = h + self.res_conv(x * mask)
|
| 114 |
+
return output
|
| 115 |
+
|
| 116 |
+
|
| 117 |
+
class Downsample1D(nn.Module):
|
| 118 |
+
def __init__(self, dim):
|
| 119 |
+
super().__init__()
|
| 120 |
+
self.conv = torch.nn.Conv1d(dim, dim, 3, 2, 1)
|
| 121 |
+
|
| 122 |
+
def forward(self, x):
|
| 123 |
+
return self.conv(x)
|
| 124 |
+
|
| 125 |
+
|
| 126 |
+
class TimestepEmbedding(nn.Module):
|
| 127 |
+
def __init__(
|
| 128 |
+
self,
|
| 129 |
+
in_channels: int,
|
| 130 |
+
time_embed_dim: int,
|
| 131 |
+
act_fn: str = "silu",
|
| 132 |
+
out_dim: int = None,
|
| 133 |
+
post_act_fn: Optional[str] = None,
|
| 134 |
+
cond_proj_dim=None,
|
| 135 |
+
):
|
| 136 |
+
super().__init__()
|
| 137 |
+
|
| 138 |
+
self.linear_1 = nn.Linear(in_channels, time_embed_dim)
|
| 139 |
+
|
| 140 |
+
if cond_proj_dim is not None:
|
| 141 |
+
self.cond_proj = nn.Linear(cond_proj_dim, in_channels, bias=False)
|
| 142 |
+
else:
|
| 143 |
+
self.cond_proj = None
|
| 144 |
+
|
| 145 |
+
self.act = get_activation(act_fn)
|
| 146 |
+
|
| 147 |
+
if out_dim is not None:
|
| 148 |
+
time_embed_dim_out = out_dim
|
| 149 |
+
else:
|
| 150 |
+
time_embed_dim_out = time_embed_dim
|
| 151 |
+
self.linear_2 = nn.Linear(time_embed_dim, time_embed_dim_out)
|
| 152 |
+
|
| 153 |
+
if post_act_fn is None:
|
| 154 |
+
self.post_act = None
|
| 155 |
+
else:
|
| 156 |
+
self.post_act = get_activation(post_act_fn)
|
| 157 |
+
|
| 158 |
+
def forward(self, sample, condition=None):
|
| 159 |
+
if condition is not None:
|
| 160 |
+
sample = sample + self.cond_proj(condition)
|
| 161 |
+
sample = self.linear_1(sample)
|
| 162 |
+
|
| 163 |
+
if self.act is not None:
|
| 164 |
+
sample = self.act(sample)
|
| 165 |
+
|
| 166 |
+
sample = self.linear_2(sample)
|
| 167 |
+
|
| 168 |
+
if self.post_act is not None:
|
| 169 |
+
sample = self.post_act(sample)
|
| 170 |
+
return sample
|
| 171 |
+
|
| 172 |
+
|
| 173 |
+
class Upsample1D(nn.Module):
|
| 174 |
+
"""A 1D upsampling layer with an optional convolution.
|
| 175 |
+
|
| 176 |
+
Parameters:
|
| 177 |
+
channels (`int`):
|
| 178 |
+
number of channels in the inputs and outputs.
|
| 179 |
+
use_conv (`bool`, default `False`):
|
| 180 |
+
option to use a convolution.
|
| 181 |
+
use_conv_transpose (`bool`, default `False`):
|
| 182 |
+
option to use a convolution transpose.
|
| 183 |
+
out_channels (`int`, optional):
|
| 184 |
+
number of output channels. Defaults to `channels`.
|
| 185 |
+
"""
|
| 186 |
+
|
| 187 |
+
def __init__(
|
| 188 |
+
self,
|
| 189 |
+
channels,
|
| 190 |
+
use_conv=False,
|
| 191 |
+
use_conv_transpose=True,
|
| 192 |
+
out_channels=None,
|
| 193 |
+
name="conv",
|
| 194 |
+
):
|
| 195 |
+
super().__init__()
|
| 196 |
+
self.channels = channels
|
| 197 |
+
self.out_channels = out_channels or channels
|
| 198 |
+
self.use_conv = use_conv
|
| 199 |
+
self.use_conv_transpose = use_conv_transpose
|
| 200 |
+
self.name = name
|
| 201 |
+
|
| 202 |
+
self.conv = None
|
| 203 |
+
if use_conv_transpose:
|
| 204 |
+
self.conv = nn.ConvTranspose1d(channels, self.out_channels, 4, 2, 1)
|
| 205 |
+
elif use_conv:
|
| 206 |
+
self.conv = nn.Conv1d(self.channels, self.out_channels, 3, padding=1)
|
| 207 |
+
|
| 208 |
+
def forward(self, inputs):
|
| 209 |
+
assert inputs.shape[1] == self.channels
|
| 210 |
+
if self.use_conv_transpose:
|
| 211 |
+
return self.conv(inputs)
|
| 212 |
+
|
| 213 |
+
outputs = F.interpolate(inputs, scale_factor=2.0, mode="nearest")
|
| 214 |
+
|
| 215 |
+
if self.use_conv:
|
| 216 |
+
outputs = self.conv(outputs)
|
| 217 |
+
|
| 218 |
+
return outputs
|
| 219 |
+
|
| 220 |
+
|
| 221 |
+
class ConformerWrapper(ConformerBlock):
|
| 222 |
+
def __init__( # pylint: disable=useless-super-delegation
|
| 223 |
+
self,
|
| 224 |
+
*,
|
| 225 |
+
dim,
|
| 226 |
+
dim_head=64,
|
| 227 |
+
heads=8,
|
| 228 |
+
ff_mult=4,
|
| 229 |
+
conv_expansion_factor=2,
|
| 230 |
+
conv_kernel_size=31,
|
| 231 |
+
attn_dropout=0,
|
| 232 |
+
ff_dropout=0,
|
| 233 |
+
conv_dropout=0,
|
| 234 |
+
conv_causal=False,
|
| 235 |
+
):
|
| 236 |
+
super().__init__(
|
| 237 |
+
dim=dim,
|
| 238 |
+
dim_head=dim_head,
|
| 239 |
+
heads=heads,
|
| 240 |
+
ff_mult=ff_mult,
|
| 241 |
+
conv_expansion_factor=conv_expansion_factor,
|
| 242 |
+
conv_kernel_size=conv_kernel_size,
|
| 243 |
+
attn_dropout=attn_dropout,
|
| 244 |
+
ff_dropout=ff_dropout,
|
| 245 |
+
conv_dropout=conv_dropout,
|
| 246 |
+
conv_causal=conv_causal,
|
| 247 |
+
)
|
| 248 |
+
|
| 249 |
+
def forward(
|
| 250 |
+
self,
|
| 251 |
+
hidden_states,
|
| 252 |
+
attention_mask,
|
| 253 |
+
encoder_hidden_states=None,
|
| 254 |
+
encoder_attention_mask=None,
|
| 255 |
+
timestep=None,
|
| 256 |
+
):
|
| 257 |
+
return super().forward(x=hidden_states, mask=attention_mask.bool())
|
| 258 |
+
|
| 259 |
+
|
| 260 |
+
class Decoder(nn.Module):
|
| 261 |
+
def __init__(
|
| 262 |
+
self,
|
| 263 |
+
in_channels,
|
| 264 |
+
out_channels,
|
| 265 |
+
channels=(256, 256),
|
| 266 |
+
dropout=0.05,
|
| 267 |
+
attention_head_dim=64,
|
| 268 |
+
n_blocks=1,
|
| 269 |
+
num_mid_blocks=2,
|
| 270 |
+
num_heads=4,
|
| 271 |
+
act_fn="snake",
|
| 272 |
+
down_block_type="transformer",
|
| 273 |
+
mid_block_type="transformer",
|
| 274 |
+
up_block_type="transformer",
|
| 275 |
+
):
|
| 276 |
+
super().__init__()
|
| 277 |
+
channels = tuple(channels)
|
| 278 |
+
self.in_channels = in_channels
|
| 279 |
+
self.out_channels = out_channels
|
| 280 |
+
|
| 281 |
+
self.time_embeddings = SinusoidalPosEmb(in_channels)
|
| 282 |
+
time_embed_dim = channels[0] * 4
|
| 283 |
+
self.time_mlp = TimestepEmbedding(
|
| 284 |
+
in_channels=in_channels,
|
| 285 |
+
time_embed_dim=time_embed_dim,
|
| 286 |
+
act_fn="silu",
|
| 287 |
+
)
|
| 288 |
+
|
| 289 |
+
self.down_blocks = nn.ModuleList([])
|
| 290 |
+
self.mid_blocks = nn.ModuleList([])
|
| 291 |
+
self.up_blocks = nn.ModuleList([])
|
| 292 |
+
|
| 293 |
+
output_channel = in_channels
|
| 294 |
+
for i in range(len(channels)): # pylint: disable=consider-using-enumerate
|
| 295 |
+
input_channel = output_channel
|
| 296 |
+
output_channel = channels[i]
|
| 297 |
+
is_last = i == len(channels) - 1
|
| 298 |
+
resnet = ResnetBlock1D(
|
| 299 |
+
dim=input_channel, dim_out=output_channel, time_emb_dim=time_embed_dim
|
| 300 |
+
)
|
| 301 |
+
transformer_blocks = nn.ModuleList(
|
| 302 |
+
[
|
| 303 |
+
self.get_block(
|
| 304 |
+
down_block_type,
|
| 305 |
+
output_channel,
|
| 306 |
+
attention_head_dim,
|
| 307 |
+
num_heads,
|
| 308 |
+
dropout,
|
| 309 |
+
act_fn,
|
| 310 |
+
)
|
| 311 |
+
for _ in range(n_blocks)
|
| 312 |
+
]
|
| 313 |
+
)
|
| 314 |
+
downsample = (
|
| 315 |
+
Downsample1D(output_channel)
|
| 316 |
+
if not is_last
|
| 317 |
+
else nn.Conv1d(output_channel, output_channel, 3, padding=1)
|
| 318 |
+
)
|
| 319 |
+
|
| 320 |
+
self.down_blocks.append(
|
| 321 |
+
nn.ModuleList([resnet, transformer_blocks, downsample])
|
| 322 |
+
)
|
| 323 |
+
|
| 324 |
+
for i in range(num_mid_blocks):
|
| 325 |
+
input_channel = channels[-1]
|
| 326 |
+
out_channels = channels[-1]
|
| 327 |
+
|
| 328 |
+
resnet = ResnetBlock1D(
|
| 329 |
+
dim=input_channel, dim_out=output_channel, time_emb_dim=time_embed_dim
|
| 330 |
+
)
|
| 331 |
+
|
| 332 |
+
transformer_blocks = nn.ModuleList(
|
| 333 |
+
[
|
| 334 |
+
self.get_block(
|
| 335 |
+
mid_block_type,
|
| 336 |
+
output_channel,
|
| 337 |
+
attention_head_dim,
|
| 338 |
+
num_heads,
|
| 339 |
+
dropout,
|
| 340 |
+
act_fn,
|
| 341 |
+
)
|
| 342 |
+
for _ in range(n_blocks)
|
| 343 |
+
]
|
| 344 |
+
)
|
| 345 |
+
|
| 346 |
+
self.mid_blocks.append(nn.ModuleList([resnet, transformer_blocks]))
|
| 347 |
+
|
| 348 |
+
channels = channels[::-1] + (channels[0],)
|
| 349 |
+
for i in range(len(channels) - 1):
|
| 350 |
+
input_channel = channels[i]
|
| 351 |
+
output_channel = channels[i + 1]
|
| 352 |
+
is_last = i == len(channels) - 2
|
| 353 |
+
|
| 354 |
+
resnet = ResnetBlock1D(
|
| 355 |
+
dim=2 * input_channel,
|
| 356 |
+
dim_out=output_channel,
|
| 357 |
+
time_emb_dim=time_embed_dim,
|
| 358 |
+
)
|
| 359 |
+
transformer_blocks = nn.ModuleList(
|
| 360 |
+
[
|
| 361 |
+
self.get_block(
|
| 362 |
+
up_block_type,
|
| 363 |
+
output_channel,
|
| 364 |
+
attention_head_dim,
|
| 365 |
+
num_heads,
|
| 366 |
+
dropout,
|
| 367 |
+
act_fn,
|
| 368 |
+
)
|
| 369 |
+
for _ in range(n_blocks)
|
| 370 |
+
]
|
| 371 |
+
)
|
| 372 |
+
upsample = (
|
| 373 |
+
Upsample1D(output_channel, use_conv_transpose=True)
|
| 374 |
+
if not is_last
|
| 375 |
+
else nn.Conv1d(output_channel, output_channel, 3, padding=1)
|
| 376 |
+
)
|
| 377 |
+
|
| 378 |
+
self.up_blocks.append(nn.ModuleList([resnet, transformer_blocks, upsample]))
|
| 379 |
+
|
| 380 |
+
self.final_block = Block1D(channels[-1], channels[-1])
|
| 381 |
+
self.final_proj = nn.Conv1d(channels[-1], self.out_channels, 1)
|
| 382 |
+
|
| 383 |
+
self.initialize_weights()
|
| 384 |
+
# nn.init.normal_(self.final_proj.weight)
|
| 385 |
+
|
| 386 |
+
@staticmethod
|
| 387 |
+
def get_block(block_type, dim, attention_head_dim, num_heads, dropout, act_fn):
|
| 388 |
+
if block_type == "conformer":
|
| 389 |
+
block = ConformerWrapper(
|
| 390 |
+
dim=dim,
|
| 391 |
+
dim_head=attention_head_dim,
|
| 392 |
+
heads=num_heads,
|
| 393 |
+
ff_mult=1,
|
| 394 |
+
conv_expansion_factor=2,
|
| 395 |
+
ff_dropout=dropout,
|
| 396 |
+
attn_dropout=dropout,
|
| 397 |
+
conv_dropout=dropout,
|
| 398 |
+
conv_kernel_size=31,
|
| 399 |
+
)
|
| 400 |
+
elif block_type == "transformer":
|
| 401 |
+
block = BasicTransformerBlock(
|
| 402 |
+
dim=dim,
|
| 403 |
+
num_attention_heads=num_heads,
|
| 404 |
+
attention_head_dim=attention_head_dim,
|
| 405 |
+
dropout=dropout,
|
| 406 |
+
activation_fn=act_fn,
|
| 407 |
+
)
|
| 408 |
+
else:
|
| 409 |
+
raise ValueError(f"Unknown block type {block_type}")
|
| 410 |
+
|
| 411 |
+
return block
|
| 412 |
+
|
| 413 |
+
def initialize_weights(self):
|
| 414 |
+
for m in self.modules():
|
| 415 |
+
if isinstance(m, nn.Conv1d):
|
| 416 |
+
nn.init.kaiming_normal_(m.weight, nonlinearity="relu")
|
| 417 |
+
|
| 418 |
+
if m.bias is not None:
|
| 419 |
+
nn.init.constant_(m.bias, 0)
|
| 420 |
+
|
| 421 |
+
elif isinstance(m, nn.GroupNorm):
|
| 422 |
+
nn.init.constant_(m.weight, 1)
|
| 423 |
+
nn.init.constant_(m.bias, 0)
|
| 424 |
+
|
| 425 |
+
elif isinstance(m, nn.Linear):
|
| 426 |
+
nn.init.kaiming_normal_(m.weight, nonlinearity="relu")
|
| 427 |
+
|
| 428 |
+
if m.bias is not None:
|
| 429 |
+
nn.init.constant_(m.bias, 0)
|
| 430 |
+
|
| 431 |
+
def forward(self, x, mask, mu, t, spks=None, cond=None):
|
| 432 |
+
"""Forward pass of the UNet1DConditional model.
|
| 433 |
+
|
| 434 |
+
Args:
|
| 435 |
+
x (torch.Tensor): shape (batch_size, in_channels, time)
|
| 436 |
+
mask (_type_): shape (batch_size, 1, time)
|
| 437 |
+
t (_type_): shape (batch_size)
|
| 438 |
+
spks (_type_, optional): shape: (batch_size, condition_channels). Defaults to None.
|
| 439 |
+
cond (_type_, optional): placeholder for future use. Defaults to None.
|
| 440 |
+
|
| 441 |
+
Raises:
|
| 442 |
+
ValueError: _description_
|
| 443 |
+
ValueError: _description_
|
| 444 |
+
|
| 445 |
+
Returns:
|
| 446 |
+
_type_: _description_
|
| 447 |
+
"""
|
| 448 |
+
|
| 449 |
+
t = self.time_embeddings(t)
|
| 450 |
+
t = self.time_mlp(t)
|
| 451 |
+
|
| 452 |
+
x = pack([x, mu], "b * t")[0]
|
| 453 |
+
|
| 454 |
+
if spks is not None:
|
| 455 |
+
spks = repeat(spks, "b c -> b c t", t=x.shape[-1])
|
| 456 |
+
x = pack([x, spks], "b * t")[0]
|
| 457 |
+
|
| 458 |
+
hiddens = []
|
| 459 |
+
masks = [mask]
|
| 460 |
+
for resnet, transformer_blocks, downsample in self.down_blocks:
|
| 461 |
+
mask_down = masks[-1]
|
| 462 |
+
x = resnet(x, mask_down, t)
|
| 463 |
+
x = rearrange(x, "b c t -> b t c")
|
| 464 |
+
mask_down = rearrange(mask_down, "b 1 t -> b t")
|
| 465 |
+
for transformer_block in transformer_blocks:
|
| 466 |
+
x = transformer_block(
|
| 467 |
+
hidden_states=x,
|
| 468 |
+
attention_mask=mask_down,
|
| 469 |
+
timestep=t,
|
| 470 |
+
)
|
| 471 |
+
x = rearrange(x, "b t c -> b c t")
|
| 472 |
+
mask_down = rearrange(mask_down, "b t -> b 1 t")
|
| 473 |
+
hiddens.append(x) # Save hidden states for skip connections
|
| 474 |
+
x = downsample(x * mask_down)
|
| 475 |
+
masks.append(mask_down[:, :, ::2])
|
| 476 |
+
|
| 477 |
+
masks = masks[:-1]
|
| 478 |
+
mask_mid = masks[-1]
|
| 479 |
+
|
| 480 |
+
for resnet, transformer_blocks in self.mid_blocks:
|
| 481 |
+
x = resnet(x, mask_mid, t)
|
| 482 |
+
x = rearrange(x, "b c t -> b t c")
|
| 483 |
+
mask_mid = rearrange(mask_mid, "b 1 t -> b t")
|
| 484 |
+
for transformer_block in transformer_blocks:
|
| 485 |
+
x = transformer_block(
|
| 486 |
+
hidden_states=x,
|
| 487 |
+
attention_mask=mask_mid,
|
| 488 |
+
timestep=t,
|
| 489 |
+
)
|
| 490 |
+
x = rearrange(x, "b t c -> b c t")
|
| 491 |
+
mask_mid = rearrange(mask_mid, "b t -> b 1 t")
|
| 492 |
+
|
| 493 |
+
for resnet, transformer_blocks, upsample in self.up_blocks:
|
| 494 |
+
mask_up = masks.pop()
|
| 495 |
+
x = resnet(pack([x, hiddens.pop()], "b * t")[0], mask_up, t)
|
| 496 |
+
x = rearrange(x, "b c t -> b t c")
|
| 497 |
+
mask_up = rearrange(mask_up, "b 1 t -> b t")
|
| 498 |
+
for transformer_block in transformer_blocks:
|
| 499 |
+
x = transformer_block(
|
| 500 |
+
hidden_states=x,
|
| 501 |
+
attention_mask=mask_up,
|
| 502 |
+
timestep=t,
|
| 503 |
+
)
|
| 504 |
+
x = rearrange(x, "b t c -> b c t")
|
| 505 |
+
mask_up = rearrange(mask_up, "b t -> b 1 t")
|
| 506 |
+
x = upsample(x * mask_up)
|
| 507 |
+
|
| 508 |
+
x = self.final_block(x, mask_up)
|
| 509 |
+
output = self.final_proj(x * mask_up)
|
| 510 |
+
|
| 511 |
+
return output * mask
|
cosyvoice/matcha/flow_matching.py
ADDED
|
@@ -0,0 +1,141 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from abc import ABC
|
| 2 |
+
|
| 3 |
+
import torch
|
| 4 |
+
import torch.nn.functional as F
|
| 5 |
+
|
| 6 |
+
from cosyvoice.matcha.decoder import Decoder
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
class BASECFM(torch.nn.Module, ABC):
|
| 10 |
+
def __init__(
|
| 11 |
+
self,
|
| 12 |
+
n_feats,
|
| 13 |
+
cfm_params,
|
| 14 |
+
n_spks=1,
|
| 15 |
+
spk_emb_dim=128,
|
| 16 |
+
):
|
| 17 |
+
super().__init__()
|
| 18 |
+
self.n_feats = n_feats
|
| 19 |
+
self.n_spks = n_spks
|
| 20 |
+
self.spk_emb_dim = spk_emb_dim
|
| 21 |
+
self.solver = cfm_params.solver
|
| 22 |
+
if hasattr(cfm_params, "sigma_min"):
|
| 23 |
+
self.sigma_min = cfm_params.sigma_min
|
| 24 |
+
else:
|
| 25 |
+
self.sigma_min = 1e-4
|
| 26 |
+
|
| 27 |
+
self.estimator = None
|
| 28 |
+
|
| 29 |
+
@torch.inference_mode()
|
| 30 |
+
def forward(self, mu, mask, n_timesteps, temperature=1.0, spks=None, cond=None):
|
| 31 |
+
"""Forward diffusion
|
| 32 |
+
|
| 33 |
+
Args:
|
| 34 |
+
mu (torch.Tensor): output of encoder
|
| 35 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 36 |
+
mask (torch.Tensor): output_mask
|
| 37 |
+
shape: (batch_size, 1, mel_timesteps)
|
| 38 |
+
n_timesteps (int): number of diffusion steps
|
| 39 |
+
temperature (float, optional): temperature for scaling noise. Defaults to 1.0.
|
| 40 |
+
spks (torch.Tensor, optional): speaker ids. Defaults to None.
|
| 41 |
+
shape: (batch_size, spk_emb_dim)
|
| 42 |
+
cond: Not used but kept for future purposes
|
| 43 |
+
|
| 44 |
+
Returns:
|
| 45 |
+
sample: generated mel-spectrogram
|
| 46 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 47 |
+
"""
|
| 48 |
+
z = torch.randn_like(mu) * temperature
|
| 49 |
+
t_span = torch.linspace(0, 1, n_timesteps + 1, device=mu.device)
|
| 50 |
+
return self.solve_euler(
|
| 51 |
+
z, t_span=t_span, mu=mu, mask=mask, spks=spks, cond=cond
|
| 52 |
+
)
|
| 53 |
+
|
| 54 |
+
def solve_euler(self, x, t_span, mu, mask, spks, cond):
|
| 55 |
+
"""
|
| 56 |
+
Fixed euler solver for ODEs.
|
| 57 |
+
Args:
|
| 58 |
+
x (torch.Tensor): random noise
|
| 59 |
+
t_span (torch.Tensor): n_timesteps interpolated
|
| 60 |
+
shape: (n_timesteps + 1,)
|
| 61 |
+
mu (torch.Tensor): output of encoder
|
| 62 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 63 |
+
mask (torch.Tensor): output_mask
|
| 64 |
+
shape: (batch_size, 1, mel_timesteps)
|
| 65 |
+
spks (torch.Tensor, optional): speaker ids. Defaults to None.
|
| 66 |
+
shape: (batch_size, spk_emb_dim)
|
| 67 |
+
cond: Not used but kept for future purposes
|
| 68 |
+
"""
|
| 69 |
+
t, _, dt = t_span[0], t_span[-1], t_span[1] - t_span[0]
|
| 70 |
+
|
| 71 |
+
# I am storing this because I can later plot it by putting a debugger here and saving it to a file
|
| 72 |
+
# Or in future might add like a return_all_steps flag
|
| 73 |
+
sol = []
|
| 74 |
+
|
| 75 |
+
for step in range(1, len(t_span)):
|
| 76 |
+
dphi_dt = self.estimator(x, mask, mu, t, spks, cond)
|
| 77 |
+
|
| 78 |
+
x = x + dt * dphi_dt
|
| 79 |
+
t = t + dt
|
| 80 |
+
sol.append(x)
|
| 81 |
+
if step < len(t_span) - 1:
|
| 82 |
+
dt = t_span[step + 1] - t
|
| 83 |
+
|
| 84 |
+
return sol[-1]
|
| 85 |
+
|
| 86 |
+
def compute_loss(self, x1, mask, mu, spks=None, cond=None):
|
| 87 |
+
"""Computes diffusion loss
|
| 88 |
+
|
| 89 |
+
Args:
|
| 90 |
+
x1 (torch.Tensor): Target
|
| 91 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 92 |
+
mask (torch.Tensor): target mask
|
| 93 |
+
shape: (batch_size, 1, mel_timesteps)
|
| 94 |
+
mu (torch.Tensor): output of encoder
|
| 95 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 96 |
+
spks (torch.Tensor, optional): speaker embedding. Defaults to None.
|
| 97 |
+
shape: (batch_size, spk_emb_dim)
|
| 98 |
+
|
| 99 |
+
Returns:
|
| 100 |
+
loss: conditional flow matching loss
|
| 101 |
+
y: conditional flow
|
| 102 |
+
shape: (batch_size, n_feats, mel_timesteps)
|
| 103 |
+
"""
|
| 104 |
+
b, _, t = mu.shape
|
| 105 |
+
|
| 106 |
+
# random timestep
|
| 107 |
+
t = torch.rand([b, 1, 1], device=mu.device, dtype=mu.dtype)
|
| 108 |
+
# sample noise p(x_0)
|
| 109 |
+
z = torch.randn_like(x1)
|
| 110 |
+
|
| 111 |
+
y = (1 - (1 - self.sigma_min) * t) * z + t * x1
|
| 112 |
+
u = x1 - (1 - self.sigma_min) * z
|
| 113 |
+
|
| 114 |
+
loss = F.mse_loss(
|
| 115 |
+
self.estimator(y, mask, mu, t.squeeze(), spks), u, reduction="sum"
|
| 116 |
+
) / (torch.sum(mask) * u.shape[1])
|
| 117 |
+
return loss, y
|
| 118 |
+
|
| 119 |
+
|
| 120 |
+
class CFM(BASECFM):
|
| 121 |
+
def __init__(
|
| 122 |
+
self,
|
| 123 |
+
in_channels,
|
| 124 |
+
out_channel,
|
| 125 |
+
cfm_params,
|
| 126 |
+
decoder_params,
|
| 127 |
+
n_spks=1,
|
| 128 |
+
spk_emb_dim=64,
|
| 129 |
+
):
|
| 130 |
+
super().__init__(
|
| 131 |
+
n_feats=in_channels,
|
| 132 |
+
cfm_params=cfm_params,
|
| 133 |
+
n_spks=n_spks,
|
| 134 |
+
spk_emb_dim=spk_emb_dim,
|
| 135 |
+
)
|
| 136 |
+
|
| 137 |
+
in_channels = in_channels + (spk_emb_dim if n_spks > 1 else 0)
|
| 138 |
+
# Just change the architecture of the estimator here
|
| 139 |
+
self.estimator = Decoder(
|
| 140 |
+
in_channels=in_channels, out_channels=out_channel, **decoder_params
|
| 141 |
+
)
|
cosyvoice/matcha/transformer.py
ADDED
|
@@ -0,0 +1,443 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from typing import Any, Dict, Optional
|
| 2 |
+
|
| 3 |
+
import torch
|
| 4 |
+
import torch.nn as nn
|
| 5 |
+
from diffusers.models.attention import (
|
| 6 |
+
GEGLU,
|
| 7 |
+
GELU,
|
| 8 |
+
AdaLayerNorm,
|
| 9 |
+
AdaLayerNormZero,
|
| 10 |
+
ApproximateGELU,
|
| 11 |
+
)
|
| 12 |
+
from diffusers.models.attention_processor import Attention
|
| 13 |
+
from diffusers.models.lora import LoRACompatibleLinear
|
| 14 |
+
from diffusers.utils.torch_utils import maybe_allow_in_graph
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
class SnakeBeta(nn.Module):
|
| 18 |
+
"""
|
| 19 |
+
A modified Snake function which uses separate parameters for the magnitude of the periodic components
|
| 20 |
+
Shape:
|
| 21 |
+
- Input: (B, C, T)
|
| 22 |
+
- Output: (B, C, T), same shape as the input
|
| 23 |
+
Parameters:
|
| 24 |
+
- alpha - trainable parameter that controls frequency
|
| 25 |
+
- beta - trainable parameter that controls magnitude
|
| 26 |
+
References:
|
| 27 |
+
- This activation function is a modified version based on this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
|
| 28 |
+
https://arxiv.org/abs/2006.08195
|
| 29 |
+
Examples:
|
| 30 |
+
>>> a1 = snakebeta(256)
|
| 31 |
+
>>> x = torch.randn(256)
|
| 32 |
+
>>> x = a1(x)
|
| 33 |
+
"""
|
| 34 |
+
|
| 35 |
+
def __init__(
|
| 36 |
+
self,
|
| 37 |
+
in_features,
|
| 38 |
+
out_features,
|
| 39 |
+
alpha=1.0,
|
| 40 |
+
alpha_trainable=True,
|
| 41 |
+
alpha_logscale=True,
|
| 42 |
+
):
|
| 43 |
+
"""
|
| 44 |
+
Initialization.
|
| 45 |
+
INPUT:
|
| 46 |
+
- in_features: shape of the input
|
| 47 |
+
- alpha - trainable parameter that controls frequency
|
| 48 |
+
- beta - trainable parameter that controls magnitude
|
| 49 |
+
alpha is initialized to 1 by default, higher values = higher-frequency.
|
| 50 |
+
beta is initialized to 1 by default, higher values = higher-magnitude.
|
| 51 |
+
alpha will be trained along with the rest of your model.
|
| 52 |
+
"""
|
| 53 |
+
super().__init__()
|
| 54 |
+
self.in_features = (
|
| 55 |
+
out_features if isinstance(out_features, list) else [out_features]
|
| 56 |
+
)
|
| 57 |
+
self.proj = LoRACompatibleLinear(in_features, out_features)
|
| 58 |
+
|
| 59 |
+
# initialize alpha
|
| 60 |
+
self.alpha_logscale = alpha_logscale
|
| 61 |
+
if self.alpha_logscale: # log scale alphas initialized to zeros
|
| 62 |
+
self.alpha = nn.Parameter(torch.zeros(self.in_features) * alpha)
|
| 63 |
+
self.beta = nn.Parameter(torch.zeros(self.in_features) * alpha)
|
| 64 |
+
else: # linear scale alphas initialized to ones
|
| 65 |
+
self.alpha = nn.Parameter(torch.ones(self.in_features) * alpha)
|
| 66 |
+
self.beta = nn.Parameter(torch.ones(self.in_features) * alpha)
|
| 67 |
+
|
| 68 |
+
self.alpha.requires_grad = alpha_trainable
|
| 69 |
+
self.beta.requires_grad = alpha_trainable
|
| 70 |
+
|
| 71 |
+
self.no_div_by_zero = 0.000000001
|
| 72 |
+
|
| 73 |
+
def forward(self, x):
|
| 74 |
+
"""
|
| 75 |
+
Forward pass of the function.
|
| 76 |
+
Applies the function to the input elementwise.
|
| 77 |
+
SnakeBeta ∶= x + 1/b * sin^2 (xa)
|
| 78 |
+
"""
|
| 79 |
+
x = self.proj(x)
|
| 80 |
+
if self.alpha_logscale:
|
| 81 |
+
alpha = torch.exp(self.alpha)
|
| 82 |
+
beta = torch.exp(self.beta)
|
| 83 |
+
else:
|
| 84 |
+
alpha = self.alpha
|
| 85 |
+
beta = self.beta
|
| 86 |
+
|
| 87 |
+
x = x + (1.0 / (beta + self.no_div_by_zero)) * torch.pow(
|
| 88 |
+
torch.sin(x * alpha), 2
|
| 89 |
+
)
|
| 90 |
+
|
| 91 |
+
return x
|
| 92 |
+
|
| 93 |
+
|
| 94 |
+
class FeedForward(nn.Module):
|
| 95 |
+
r"""
|
| 96 |
+
A feed-forward layer.
|
| 97 |
+
|
| 98 |
+
Parameters:
|
| 99 |
+
dim (`int`): The number of channels in the input.
|
| 100 |
+
dim_out (`int`, *optional*): The number of channels in the output. If not given, defaults to `dim`.
|
| 101 |
+
mult (`int`, *optional*, defaults to 4): The multiplier to use for the hidden dimension.
|
| 102 |
+
dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use.
|
| 103 |
+
activation_fn (`str`, *optional*, defaults to `"geglu"`): Activation function to be used in feed-forward.
|
| 104 |
+
final_dropout (`bool` *optional*, defaults to False): Apply a final dropout.
|
| 105 |
+
"""
|
| 106 |
+
|
| 107 |
+
def __init__(
|
| 108 |
+
self,
|
| 109 |
+
dim: int,
|
| 110 |
+
dim_out: Optional[int] = None,
|
| 111 |
+
mult: int = 4,
|
| 112 |
+
dropout: float = 0.0,
|
| 113 |
+
activation_fn: str = "geglu",
|
| 114 |
+
final_dropout: bool = False,
|
| 115 |
+
):
|
| 116 |
+
super().__init__()
|
| 117 |
+
inner_dim = int(dim * mult)
|
| 118 |
+
dim_out = dim_out if dim_out is not None else dim
|
| 119 |
+
|
| 120 |
+
if activation_fn == "gelu":
|
| 121 |
+
act_fn = GELU(dim, inner_dim)
|
| 122 |
+
if activation_fn == "gelu-approximate":
|
| 123 |
+
act_fn = GELU(dim, inner_dim, approximate="tanh")
|
| 124 |
+
elif activation_fn == "geglu":
|
| 125 |
+
act_fn = GEGLU(dim, inner_dim)
|
| 126 |
+
elif activation_fn == "geglu-approximate":
|
| 127 |
+
act_fn = ApproximateGELU(dim, inner_dim)
|
| 128 |
+
elif activation_fn == "snakebeta":
|
| 129 |
+
act_fn = SnakeBeta(dim, inner_dim)
|
| 130 |
+
|
| 131 |
+
self.net = nn.ModuleList([])
|
| 132 |
+
# project in
|
| 133 |
+
self.net.append(act_fn)
|
| 134 |
+
# project dropout
|
| 135 |
+
self.net.append(nn.Dropout(dropout))
|
| 136 |
+
# project out
|
| 137 |
+
self.net.append(LoRACompatibleLinear(inner_dim, dim_out))
|
| 138 |
+
# FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout
|
| 139 |
+
if final_dropout:
|
| 140 |
+
self.net.append(nn.Dropout(dropout))
|
| 141 |
+
|
| 142 |
+
def forward(self, hidden_states):
|
| 143 |
+
for module in self.net:
|
| 144 |
+
hidden_states = module(hidden_states)
|
| 145 |
+
return hidden_states
|
| 146 |
+
|
| 147 |
+
|
| 148 |
+
@maybe_allow_in_graph
|
| 149 |
+
class BasicTransformerBlock(nn.Module):
|
| 150 |
+
r"""
|
| 151 |
+
A basic Transformer block.
|
| 152 |
+
|
| 153 |
+
Parameters:
|
| 154 |
+
dim (`int`): The number of channels in the input and output.
|
| 155 |
+
num_attention_heads (`int`): The number of heads to use for multi-head attention.
|
| 156 |
+
attention_head_dim (`int`): The number of channels in each head.
|
| 157 |
+
dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use.
|
| 158 |
+
cross_attention_dim (`int`, *optional*): The size of the encoder_hidden_states vector for cross attention.
|
| 159 |
+
only_cross_attention (`bool`, *optional*):
|
| 160 |
+
Whether to use only cross-attention layers. In this case two cross attention layers are used.
|
| 161 |
+
double_self_attention (`bool`, *optional*):
|
| 162 |
+
Whether to use two self-attention layers. In this case no cross attention layers are used.
|
| 163 |
+
activation_fn (`str`, *optional*, defaults to `"geglu"`): Activation function to be used in feed-forward.
|
| 164 |
+
num_embeds_ada_norm (:
|
| 165 |
+
obj: `int`, *optional*): The number of diffusion steps used during training. See `Transformer2DModel`.
|
| 166 |
+
attention_bias (:
|
| 167 |
+
obj: `bool`, *optional*, defaults to `False`): Configure if the attentions should contain a bias parameter.
|
| 168 |
+
"""
|
| 169 |
+
|
| 170 |
+
def __init__(
|
| 171 |
+
self,
|
| 172 |
+
dim: int,
|
| 173 |
+
num_attention_heads: int,
|
| 174 |
+
attention_head_dim: int,
|
| 175 |
+
dropout=0.0,
|
| 176 |
+
cross_attention_dim: Optional[int] = None,
|
| 177 |
+
activation_fn: str = "geglu",
|
| 178 |
+
num_embeds_ada_norm: Optional[int] = None,
|
| 179 |
+
attention_bias: bool = False,
|
| 180 |
+
only_cross_attention: bool = False,
|
| 181 |
+
double_self_attention: bool = False,
|
| 182 |
+
upcast_attention: bool = False,
|
| 183 |
+
norm_elementwise_affine: bool = True,
|
| 184 |
+
norm_type: str = "layer_norm",
|
| 185 |
+
final_dropout: bool = False,
|
| 186 |
+
):
|
| 187 |
+
super().__init__()
|
| 188 |
+
self.only_cross_attention = only_cross_attention
|
| 189 |
+
|
| 190 |
+
self.use_ada_layer_norm_zero = (
|
| 191 |
+
num_embeds_ada_norm is not None
|
| 192 |
+
) and norm_type == "ada_norm_zero"
|
| 193 |
+
self.use_ada_layer_norm = (
|
| 194 |
+
num_embeds_ada_norm is not None
|
| 195 |
+
) and norm_type == "ada_norm"
|
| 196 |
+
|
| 197 |
+
if norm_type in ("ada_norm", "ada_norm_zero") and num_embeds_ada_norm is None:
|
| 198 |
+
raise ValueError(
|
| 199 |
+
f"`norm_type` is set to {norm_type}, but `num_embeds_ada_norm` is not defined. Please make sure to"
|
| 200 |
+
f" define `num_embeds_ada_norm` if setting `norm_type` to {norm_type}."
|
| 201 |
+
)
|
| 202 |
+
|
| 203 |
+
# Define 3 blocks. Each block has its own normalization layer.
|
| 204 |
+
# 1. Self-Attn
|
| 205 |
+
if self.use_ada_layer_norm:
|
| 206 |
+
self.norm1 = AdaLayerNorm(dim, num_embeds_ada_norm)
|
| 207 |
+
elif self.use_ada_layer_norm_zero:
|
| 208 |
+
self.norm1 = AdaLayerNormZero(dim, num_embeds_ada_norm)
|
| 209 |
+
else:
|
| 210 |
+
self.norm1 = nn.LayerNorm(dim, elementwise_affine=norm_elementwise_affine)
|
| 211 |
+
self.attn1 = Attention(
|
| 212 |
+
query_dim=dim,
|
| 213 |
+
heads=num_attention_heads,
|
| 214 |
+
dim_head=attention_head_dim,
|
| 215 |
+
dropout=dropout,
|
| 216 |
+
bias=attention_bias,
|
| 217 |
+
cross_attention_dim=cross_attention_dim if only_cross_attention else None,
|
| 218 |
+
upcast_attention=upcast_attention,
|
| 219 |
+
)
|
| 220 |
+
|
| 221 |
+
# 2. Cross-Attn
|
| 222 |
+
if cross_attention_dim is not None or double_self_attention:
|
| 223 |
+
# We currently only use AdaLayerNormZero for self attention where there will only be one attention block.
|
| 224 |
+
# I.e. the number of returned modulation chunks from AdaLayerZero would not make sense if returned during
|
| 225 |
+
# the second cross attention block.
|
| 226 |
+
self.norm2 = (
|
| 227 |
+
AdaLayerNorm(dim, num_embeds_ada_norm)
|
| 228 |
+
if self.use_ada_layer_norm
|
| 229 |
+
else nn.LayerNorm(dim, elementwise_affine=norm_elementwise_affine)
|
| 230 |
+
)
|
| 231 |
+
self.attn2 = Attention(
|
| 232 |
+
query_dim=dim,
|
| 233 |
+
cross_attention_dim=(
|
| 234 |
+
cross_attention_dim if not double_self_attention else None
|
| 235 |
+
),
|
| 236 |
+
heads=num_attention_heads,
|
| 237 |
+
dim_head=attention_head_dim,
|
| 238 |
+
dropout=dropout,
|
| 239 |
+
bias=attention_bias,
|
| 240 |
+
upcast_attention=upcast_attention,
|
| 241 |
+
# scale_qk=False, # uncomment this to not to use flash attention
|
| 242 |
+
) # is self-attn if encoder_hidden_states is none
|
| 243 |
+
else:
|
| 244 |
+
self.norm2 = None
|
| 245 |
+
self.attn2 = None
|
| 246 |
+
|
| 247 |
+
# 3. Feed-forward
|
| 248 |
+
self.norm3 = nn.LayerNorm(dim, elementwise_affine=norm_elementwise_affine)
|
| 249 |
+
self.ff = FeedForward(
|
| 250 |
+
dim,
|
| 251 |
+
dropout=dropout,
|
| 252 |
+
activation_fn=activation_fn,
|
| 253 |
+
final_dropout=final_dropout,
|
| 254 |
+
)
|
| 255 |
+
|
| 256 |
+
# let chunk size default to None
|
| 257 |
+
self._chunk_size = None
|
| 258 |
+
self._chunk_dim = 0
|
| 259 |
+
|
| 260 |
+
def set_chunk_feed_forward(self, chunk_size: Optional[int], dim: int):
|
| 261 |
+
# Sets chunk feed-forward
|
| 262 |
+
self._chunk_size = chunk_size
|
| 263 |
+
self._chunk_dim = dim
|
| 264 |
+
|
| 265 |
+
def forward_native(
|
| 266 |
+
self,
|
| 267 |
+
hidden_states: torch.FloatTensor,
|
| 268 |
+
attention_mask: Optional[torch.FloatTensor] = None,
|
| 269 |
+
encoder_hidden_states: Optional[torch.FloatTensor] = None,
|
| 270 |
+
encoder_attention_mask: Optional[torch.FloatTensor] = None,
|
| 271 |
+
timestep: Optional[torch.LongTensor] = None,
|
| 272 |
+
cross_attention_kwargs: Dict[str, Any] = None,
|
| 273 |
+
class_labels: Optional[torch.LongTensor] = None,
|
| 274 |
+
):
|
| 275 |
+
# Notice that normalization is always applied before the real computation in the following blocks.
|
| 276 |
+
# 1. Self-Attention
|
| 277 |
+
if self.use_ada_layer_norm:
|
| 278 |
+
norm_hidden_states = self.norm1(hidden_states, timestep)
|
| 279 |
+
elif self.use_ada_layer_norm_zero:
|
| 280 |
+
norm_hidden_states, gate_msa, shift_mlp, scale_mlp, gate_mlp = self.norm1(
|
| 281 |
+
hidden_states, timestep, class_labels, hidden_dtype=hidden_states.dtype
|
| 282 |
+
)
|
| 283 |
+
else:
|
| 284 |
+
norm_hidden_states = self.norm1(hidden_states)
|
| 285 |
+
|
| 286 |
+
cross_attention_kwargs = (
|
| 287 |
+
cross_attention_kwargs if cross_attention_kwargs is not None else {}
|
| 288 |
+
)
|
| 289 |
+
|
| 290 |
+
attn_output = self.attn1(
|
| 291 |
+
norm_hidden_states,
|
| 292 |
+
encoder_hidden_states=(
|
| 293 |
+
encoder_hidden_states if self.only_cross_attention else None
|
| 294 |
+
),
|
| 295 |
+
attention_mask=(
|
| 296 |
+
encoder_attention_mask if self.only_cross_attention else attention_mask
|
| 297 |
+
),
|
| 298 |
+
**cross_attention_kwargs,
|
| 299 |
+
)
|
| 300 |
+
if self.use_ada_layer_norm_zero:
|
| 301 |
+
attn_output = gate_msa.unsqueeze(1) * attn_output
|
| 302 |
+
hidden_states = attn_output + hidden_states
|
| 303 |
+
|
| 304 |
+
# 2. Cross-Attention
|
| 305 |
+
if self.attn2 is not None:
|
| 306 |
+
norm_hidden_states = (
|
| 307 |
+
self.norm2(hidden_states, timestep)
|
| 308 |
+
if self.use_ada_layer_norm
|
| 309 |
+
else self.norm2(hidden_states)
|
| 310 |
+
)
|
| 311 |
+
|
| 312 |
+
attn_output = self.attn2(
|
| 313 |
+
norm_hidden_states,
|
| 314 |
+
encoder_hidden_states=encoder_hidden_states,
|
| 315 |
+
attention_mask=encoder_attention_mask,
|
| 316 |
+
**cross_attention_kwargs,
|
| 317 |
+
)
|
| 318 |
+
hidden_states = attn_output + hidden_states
|
| 319 |
+
|
| 320 |
+
# 3. Feed-forward
|
| 321 |
+
norm_hidden_states = self.norm3(hidden_states)
|
| 322 |
+
|
| 323 |
+
if self.use_ada_layer_norm_zero:
|
| 324 |
+
norm_hidden_states = (
|
| 325 |
+
norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None]
|
| 326 |
+
)
|
| 327 |
+
|
| 328 |
+
if self._chunk_size is not None:
|
| 329 |
+
# "feed_forward_chunk_size" can be used to save memory
|
| 330 |
+
if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0:
|
| 331 |
+
raise ValueError(
|
| 332 |
+
f"`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`."
|
| 333 |
+
)
|
| 334 |
+
|
| 335 |
+
num_chunks = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size
|
| 336 |
+
ff_output = torch.cat(
|
| 337 |
+
[
|
| 338 |
+
self.ff(hid_slice)
|
| 339 |
+
for hid_slice in norm_hidden_states.chunk(
|
| 340 |
+
num_chunks, dim=self._chunk_dim
|
| 341 |
+
)
|
| 342 |
+
],
|
| 343 |
+
dim=self._chunk_dim,
|
| 344 |
+
)
|
| 345 |
+
else:
|
| 346 |
+
ff_output = self.ff(norm_hidden_states)
|
| 347 |
+
|
| 348 |
+
if self.use_ada_layer_norm_zero:
|
| 349 |
+
ff_output = gate_mlp.unsqueeze(1) * ff_output
|
| 350 |
+
|
| 351 |
+
hidden_states = ff_output + hidden_states
|
| 352 |
+
|
| 353 |
+
return hidden_states
|
| 354 |
+
|
| 355 |
+
def forward(
|
| 356 |
+
self,
|
| 357 |
+
hidden_states: torch.FloatTensor,
|
| 358 |
+
attention_mask: Optional[torch.FloatTensor] = None,
|
| 359 |
+
encoder_hidden_states: Optional[torch.FloatTensor] = None,
|
| 360 |
+
encoder_attention_mask: Optional[torch.FloatTensor] = None,
|
| 361 |
+
timestep: Optional[torch.LongTensor] = None,
|
| 362 |
+
cross_attention_kwargs: Dict[str, Any] = None,
|
| 363 |
+
class_labels: Optional[torch.LongTensor] = None,
|
| 364 |
+
):
|
| 365 |
+
# Notice that normalization is always applied before the real computation in the following blocks.
|
| 366 |
+
# 1. Self-Attention
|
| 367 |
+
if self.use_ada_layer_norm:
|
| 368 |
+
norm_hidden_states = self.norm1(hidden_states, timestep)
|
| 369 |
+
elif self.use_ada_layer_norm_zero:
|
| 370 |
+
norm_hidden_states, gate_msa, shift_mlp, scale_mlp, gate_mlp = self.norm1(
|
| 371 |
+
hidden_states, timestep, class_labels, hidden_dtype=hidden_states.dtype
|
| 372 |
+
)
|
| 373 |
+
else:
|
| 374 |
+
norm_hidden_states = self.norm1(hidden_states)
|
| 375 |
+
|
| 376 |
+
cross_attention_kwargs = (
|
| 377 |
+
cross_attention_kwargs if cross_attention_kwargs is not None else {}
|
| 378 |
+
)
|
| 379 |
+
|
| 380 |
+
attn_output = self.attn1(
|
| 381 |
+
norm_hidden_states,
|
| 382 |
+
encoder_hidden_states=(
|
| 383 |
+
encoder_hidden_states if self.only_cross_attention else None
|
| 384 |
+
),
|
| 385 |
+
attention_mask=(
|
| 386 |
+
encoder_attention_mask if self.only_cross_attention else attention_mask
|
| 387 |
+
),
|
| 388 |
+
**cross_attention_kwargs,
|
| 389 |
+
)
|
| 390 |
+
if self.use_ada_layer_norm_zero:
|
| 391 |
+
attn_output = gate_msa.unsqueeze(1) * attn_output
|
| 392 |
+
hidden_states = attn_output + hidden_states
|
| 393 |
+
|
| 394 |
+
# 2. Cross-Attention
|
| 395 |
+
if self.attn2 is not None:
|
| 396 |
+
norm_hidden_states = (
|
| 397 |
+
self.norm2(hidden_states, timestep)
|
| 398 |
+
if self.use_ada_layer_norm
|
| 399 |
+
else self.norm2(hidden_states)
|
| 400 |
+
)
|
| 401 |
+
|
| 402 |
+
attn_output = self.attn2(
|
| 403 |
+
norm_hidden_states,
|
| 404 |
+
encoder_hidden_states=encoder_hidden_states,
|
| 405 |
+
attention_mask=encoder_attention_mask,
|
| 406 |
+
**cross_attention_kwargs,
|
| 407 |
+
)
|
| 408 |
+
hidden_states = attn_output + hidden_states
|
| 409 |
+
|
| 410 |
+
# 3. Feed-forward
|
| 411 |
+
norm_hidden_states = self.norm3(hidden_states)
|
| 412 |
+
|
| 413 |
+
if self.use_ada_layer_norm_zero:
|
| 414 |
+
norm_hidden_states = (
|
| 415 |
+
norm_hidden_states * (1 + scale_mlp[:, None]) + shift_mlp[:, None]
|
| 416 |
+
)
|
| 417 |
+
|
| 418 |
+
if self._chunk_size is not None:
|
| 419 |
+
# "feed_forward_chunk_size" can be used to save memory
|
| 420 |
+
if norm_hidden_states.shape[self._chunk_dim] % self._chunk_size != 0:
|
| 421 |
+
raise ValueError(
|
| 422 |
+
f"`hidden_states` dimension to be chunked: {norm_hidden_states.shape[self._chunk_dim]} has to be divisible by chunk size: {self._chunk_size}. Make sure to set an appropriate `chunk_size` when calling `unet.enable_forward_chunking`."
|
| 423 |
+
)
|
| 424 |
+
|
| 425 |
+
num_chunks = norm_hidden_states.shape[self._chunk_dim] // self._chunk_size
|
| 426 |
+
ff_output = torch.cat(
|
| 427 |
+
[
|
| 428 |
+
self.ff(hid_slice)
|
| 429 |
+
for hid_slice in norm_hidden_states.chunk(
|
| 430 |
+
num_chunks, dim=self._chunk_dim
|
| 431 |
+
)
|
| 432 |
+
],
|
| 433 |
+
dim=self._chunk_dim,
|
| 434 |
+
)
|
| 435 |
+
else:
|
| 436 |
+
ff_output = self.ff(norm_hidden_states)
|
| 437 |
+
|
| 438 |
+
if self.use_ada_layer_norm_zero:
|
| 439 |
+
ff_output = gate_mlp.unsqueeze(1) * ff_output
|
| 440 |
+
|
| 441 |
+
hidden_states = ff_output + hidden_states
|
| 442 |
+
|
| 443 |
+
return hidden_states
|
cosyvoice/transformer/__init__.py
ADDED
|
File without changes
|
cosyvoice/transformer/activation.py
ADDED
|
@@ -0,0 +1,87 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2020 Johns Hopkins University (Shinji Watanabe)
|
| 2 |
+
# 2020 Northwestern Polytechnical University (Pengcheng Guo)
|
| 3 |
+
# 2020 Mobvoi Inc (Binbin Zhang)
|
| 4 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 5 |
+
#
|
| 6 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 7 |
+
# you may not use this file except in compliance with the License.
|
| 8 |
+
# You may obtain a copy of the License at
|
| 9 |
+
#
|
| 10 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 11 |
+
#
|
| 12 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 13 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 14 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 15 |
+
# See the License for the specific language governing permissions and
|
| 16 |
+
# limitations under the License.
|
| 17 |
+
"""Swish() activation function for Conformer."""
|
| 18 |
+
|
| 19 |
+
import torch
|
| 20 |
+
from torch import nn, sin, pow
|
| 21 |
+
from torch.nn import Parameter
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
class Swish(torch.nn.Module):
|
| 25 |
+
"""Construct an Swish object."""
|
| 26 |
+
|
| 27 |
+
def forward(self, x: torch.Tensor) -> torch.Tensor:
|
| 28 |
+
"""Return Swish activation function."""
|
| 29 |
+
return x * torch.sigmoid(x)
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
# Implementation adapted from https://github.com/EdwardDixon/snake under the MIT license.
|
| 33 |
+
# LICENSE is in incl_licenses directory.
|
| 34 |
+
class Snake(nn.Module):
|
| 35 |
+
"""
|
| 36 |
+
Implementation of a sine-based periodic activation function
|
| 37 |
+
Shape:
|
| 38 |
+
- Input: (B, C, T)
|
| 39 |
+
- Output: (B, C, T), same shape as the input
|
| 40 |
+
Parameters:
|
| 41 |
+
- alpha - trainable parameter
|
| 42 |
+
References:
|
| 43 |
+
- This activation function is from this paper by Liu Ziyin, Tilman Hartwig, Masahito Ueda:
|
| 44 |
+
https://arxiv.org/abs/2006.08195
|
| 45 |
+
Examples:
|
| 46 |
+
>>> a1 = snake(256)
|
| 47 |
+
>>> x = torch.randn(256)
|
| 48 |
+
>>> x = a1(x)
|
| 49 |
+
"""
|
| 50 |
+
|
| 51 |
+
def __init__(
|
| 52 |
+
self, in_features, alpha=1.0, alpha_trainable=True, alpha_logscale=False
|
| 53 |
+
):
|
| 54 |
+
"""
|
| 55 |
+
Initialization.
|
| 56 |
+
INPUT:
|
| 57 |
+
- in_features: shape of the input
|
| 58 |
+
- alpha: trainable parameter
|
| 59 |
+
alpha is initialized to 1 by default, higher values = higher-frequency.
|
| 60 |
+
alpha will be trained along with the rest of your model.
|
| 61 |
+
"""
|
| 62 |
+
super(Snake, self).__init__()
|
| 63 |
+
self.in_features = in_features
|
| 64 |
+
|
| 65 |
+
# initialize alpha
|
| 66 |
+
self.alpha_logscale = alpha_logscale
|
| 67 |
+
if self.alpha_logscale: # log scale alphas initialized to zeros
|
| 68 |
+
self.alpha = Parameter(torch.zeros(in_features) * alpha)
|
| 69 |
+
else: # linear scale alphas initialized to ones
|
| 70 |
+
self.alpha = Parameter(torch.ones(in_features) * alpha)
|
| 71 |
+
|
| 72 |
+
self.alpha.requires_grad = alpha_trainable
|
| 73 |
+
|
| 74 |
+
self.no_div_by_zero = 0.000000001
|
| 75 |
+
|
| 76 |
+
def forward(self, x):
|
| 77 |
+
"""
|
| 78 |
+
Forward pass of the function.
|
| 79 |
+
Applies the function to the input elementwise.
|
| 80 |
+
Snake ∶= x + 1/a * sin^2 (xa)
|
| 81 |
+
"""
|
| 82 |
+
alpha = self.alpha.unsqueeze(0).unsqueeze(-1) # line up with x to [B, C, T]
|
| 83 |
+
if self.alpha_logscale:
|
| 84 |
+
alpha = torch.exp(alpha)
|
| 85 |
+
x = x + (1.0 / (alpha + self.no_div_by_zero)) * pow(sin(x * alpha), 2)
|
| 86 |
+
|
| 87 |
+
return x
|
cosyvoice/transformer/attention.py
ADDED
|
@@ -0,0 +1,322 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2019 Shigeki Karita
|
| 2 |
+
# 2020 Mobvoi Inc (Binbin Zhang)
|
| 3 |
+
# 2022 Xingchen Song ([email protected])
|
| 4 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 5 |
+
#
|
| 6 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 7 |
+
# you may not use this file except in compliance with the License.
|
| 8 |
+
# You may obtain a copy of the License at
|
| 9 |
+
#
|
| 10 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 11 |
+
#
|
| 12 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 13 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 14 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 15 |
+
# See the License for the specific language governing permissions and
|
| 16 |
+
# limitations under the License.
|
| 17 |
+
"""Multi-Head Attention layer definition."""
|
| 18 |
+
|
| 19 |
+
import math
|
| 20 |
+
from typing import Tuple
|
| 21 |
+
|
| 22 |
+
import torch
|
| 23 |
+
from torch import nn
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
class MultiHeadedAttention(nn.Module):
|
| 27 |
+
"""Multi-Head Attention layer.
|
| 28 |
+
|
| 29 |
+
Args:
|
| 30 |
+
n_head (int): The number of heads.
|
| 31 |
+
n_feat (int): The number of features.
|
| 32 |
+
dropout_rate (float): Dropout rate.
|
| 33 |
+
|
| 34 |
+
"""
|
| 35 |
+
|
| 36 |
+
def __init__(
|
| 37 |
+
self, n_head: int, n_feat: int, dropout_rate: float, key_bias: bool = True
|
| 38 |
+
):
|
| 39 |
+
"""Construct an MultiHeadedAttention object."""
|
| 40 |
+
super().__init__()
|
| 41 |
+
assert n_feat % n_head == 0
|
| 42 |
+
# We assume d_v always equals d_k
|
| 43 |
+
self.d_k = n_feat // n_head
|
| 44 |
+
self.h = n_head
|
| 45 |
+
self.linear_q = nn.Linear(n_feat, n_feat)
|
| 46 |
+
self.linear_k = nn.Linear(n_feat, n_feat, bias=key_bias)
|
| 47 |
+
self.linear_v = nn.Linear(n_feat, n_feat)
|
| 48 |
+
self.linear_out = nn.Linear(n_feat, n_feat)
|
| 49 |
+
self.dropout = nn.Dropout(p=dropout_rate)
|
| 50 |
+
|
| 51 |
+
def forward_qkv(
|
| 52 |
+
self, query: torch.Tensor, key: torch.Tensor, value: torch.Tensor
|
| 53 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 54 |
+
"""Transform query, key and value.
|
| 55 |
+
|
| 56 |
+
Args:
|
| 57 |
+
query (torch.Tensor): Query tensor (#batch, time1, size).
|
| 58 |
+
key (torch.Tensor): Key tensor (#batch, time2, size).
|
| 59 |
+
value (torch.Tensor): Value tensor (#batch, time2, size).
|
| 60 |
+
|
| 61 |
+
Returns:
|
| 62 |
+
torch.Tensor: Transformed query tensor, size
|
| 63 |
+
(#batch, n_head, time1, d_k).
|
| 64 |
+
torch.Tensor: Transformed key tensor, size
|
| 65 |
+
(#batch, n_head, time2, d_k).
|
| 66 |
+
torch.Tensor: Transformed value tensor, size
|
| 67 |
+
(#batch, n_head, time2, d_k).
|
| 68 |
+
|
| 69 |
+
"""
|
| 70 |
+
n_batch = query.size(0)
|
| 71 |
+
q = self.linear_q(query).view(n_batch, -1, self.h, self.d_k)
|
| 72 |
+
k = self.linear_k(key).view(n_batch, -1, self.h, self.d_k)
|
| 73 |
+
v = self.linear_v(value).view(n_batch, -1, self.h, self.d_k)
|
| 74 |
+
q = q.transpose(1, 2) # (batch, head, time1, d_k)
|
| 75 |
+
k = k.transpose(1, 2) # (batch, head, time2, d_k)
|
| 76 |
+
v = v.transpose(1, 2) # (batch, head, time2, d_k)
|
| 77 |
+
|
| 78 |
+
return q, k, v
|
| 79 |
+
|
| 80 |
+
def forward_attention(
|
| 81 |
+
self,
|
| 82 |
+
value: torch.Tensor,
|
| 83 |
+
scores: torch.Tensor,
|
| 84 |
+
mask: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 85 |
+
) -> torch.Tensor:
|
| 86 |
+
"""Compute attention context vector.
|
| 87 |
+
|
| 88 |
+
Args:
|
| 89 |
+
value (torch.Tensor): Transformed value, size
|
| 90 |
+
(#batch, n_head, time2, d_k).
|
| 91 |
+
scores (torch.Tensor): Attention score, size
|
| 92 |
+
(#batch, n_head, time1, time2).
|
| 93 |
+
mask (torch.Tensor): Mask, size (#batch, 1, time2) or
|
| 94 |
+
(#batch, time1, time2), (0, 0, 0) means fake mask.
|
| 95 |
+
|
| 96 |
+
Returns:
|
| 97 |
+
torch.Tensor: Transformed value (#batch, time1, d_model)
|
| 98 |
+
weighted by the attention score (#batch, time1, time2).
|
| 99 |
+
|
| 100 |
+
"""
|
| 101 |
+
n_batch = value.size(0)
|
| 102 |
+
# NOTE(xcsong): When will `if mask.size(2) > 0` be True?
|
| 103 |
+
# 1. onnx(16/4) [WHY? Because we feed real cache & real mask for the
|
| 104 |
+
# 1st chunk to ease the onnx export.]
|
| 105 |
+
# 2. pytorch training
|
| 106 |
+
if mask.size(2) > 0: # time2 > 0
|
| 107 |
+
mask = mask.unsqueeze(1).eq(0) # (batch, 1, *, time2)
|
| 108 |
+
# For last chunk, time2 might be larger than scores.size(-1)
|
| 109 |
+
mask = mask[:, :, :, : scores.size(-1)] # (batch, 1, *, time2)
|
| 110 |
+
scores = scores.masked_fill(mask, -float("inf"))
|
| 111 |
+
attn = torch.softmax(scores, dim=-1).masked_fill(
|
| 112 |
+
mask, 0.0
|
| 113 |
+
) # (batch, head, time1, time2)
|
| 114 |
+
# NOTE(xcsong): When will `if mask.size(2) > 0` be False?
|
| 115 |
+
# 1. onnx(16/-1, -1/-1, 16/0)
|
| 116 |
+
# 2. jit (16/-1, -1/-1, 16/0, 16/4)
|
| 117 |
+
else:
|
| 118 |
+
attn = torch.softmax(scores, dim=-1) # (batch, head, time1, time2)
|
| 119 |
+
|
| 120 |
+
p_attn = self.dropout(attn)
|
| 121 |
+
x = torch.matmul(p_attn, value) # (batch, head, time1, d_k)
|
| 122 |
+
x = (
|
| 123 |
+
x.transpose(1, 2).contiguous().view(n_batch, -1, self.h * self.d_k)
|
| 124 |
+
) # (batch, time1, d_model)
|
| 125 |
+
|
| 126 |
+
return self.linear_out(x) # (batch, time1, d_model)
|
| 127 |
+
|
| 128 |
+
def forward(
|
| 129 |
+
self,
|
| 130 |
+
query: torch.Tensor,
|
| 131 |
+
key: torch.Tensor,
|
| 132 |
+
value: torch.Tensor,
|
| 133 |
+
mask: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 134 |
+
pos_emb: torch.Tensor = torch.empty(0),
|
| 135 |
+
cache: torch.Tensor = torch.zeros((0, 0, 0, 0)),
|
| 136 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 137 |
+
"""Compute scaled dot product attention.
|
| 138 |
+
|
| 139 |
+
Args:
|
| 140 |
+
query (torch.Tensor): Query tensor (#batch, time1, size).
|
| 141 |
+
key (torch.Tensor): Key tensor (#batch, time2, size).
|
| 142 |
+
value (torch.Tensor): Value tensor (#batch, time2, size).
|
| 143 |
+
mask (torch.Tensor): Mask tensor (#batch, 1, time2) or
|
| 144 |
+
(#batch, time1, time2).
|
| 145 |
+
1.When applying cross attention between decoder and encoder,
|
| 146 |
+
the batch padding mask for input is in (#batch, 1, T) shape.
|
| 147 |
+
2.When applying self attention of encoder,
|
| 148 |
+
the mask is in (#batch, T, T) shape.
|
| 149 |
+
3.When applying self attention of decoder,
|
| 150 |
+
the mask is in (#batch, L, L) shape.
|
| 151 |
+
4.If the different position in decoder see different block
|
| 152 |
+
of the encoder, such as Mocha, the passed in mask could be
|
| 153 |
+
in (#batch, L, T) shape. But there is no such case in current
|
| 154 |
+
CosyVoice.
|
| 155 |
+
cache (torch.Tensor): Cache tensor (1, head, cache_t, d_k * 2),
|
| 156 |
+
where `cache_t == chunk_size * num_decoding_left_chunks`
|
| 157 |
+
and `head * d_k == size`
|
| 158 |
+
|
| 159 |
+
|
| 160 |
+
Returns:
|
| 161 |
+
torch.Tensor: Output tensor (#batch, time1, d_model).
|
| 162 |
+
torch.Tensor: Cache tensor (1, head, cache_t + time1, d_k * 2)
|
| 163 |
+
where `cache_t == chunk_size * num_decoding_left_chunks`
|
| 164 |
+
and `head * d_k == size`
|
| 165 |
+
|
| 166 |
+
"""
|
| 167 |
+
q, k, v = self.forward_qkv(query, key, value)
|
| 168 |
+
|
| 169 |
+
# NOTE(xcsong):
|
| 170 |
+
# when export onnx model, for 1st chunk, we feed
|
| 171 |
+
# cache(1, head, 0, d_k * 2) (16/-1, -1/-1, 16/0 mode)
|
| 172 |
+
# or cache(1, head, real_cache_t, d_k * 2) (16/4 mode).
|
| 173 |
+
# In all modes, `if cache.size(0) > 0` will alwayse be `True`
|
| 174 |
+
# and we will always do splitting and
|
| 175 |
+
# concatnation(this will simplify onnx export). Note that
|
| 176 |
+
# it's OK to concat & split zero-shaped tensors(see code below).
|
| 177 |
+
# when export jit model, for 1st chunk, we always feed
|
| 178 |
+
# cache(0, 0, 0, 0) since jit supports dynamic if-branch.
|
| 179 |
+
# >>> a = torch.ones((1, 2, 0, 4))
|
| 180 |
+
# >>> b = torch.ones((1, 2, 3, 4))
|
| 181 |
+
# >>> c = torch.cat((a, b), dim=2)
|
| 182 |
+
# >>> torch.equal(b, c) # True
|
| 183 |
+
# >>> d = torch.split(a, 2, dim=-1)
|
| 184 |
+
# >>> torch.equal(d[0], d[1]) # True
|
| 185 |
+
if cache.size(0) > 0:
|
| 186 |
+
key_cache, value_cache = torch.split(cache, cache.size(-1) // 2, dim=-1)
|
| 187 |
+
k = torch.cat([key_cache, k], dim=2)
|
| 188 |
+
v = torch.cat([value_cache, v], dim=2)
|
| 189 |
+
# NOTE(xcsong): We do cache slicing in encoder.forward_chunk, since it's
|
| 190 |
+
# non-trivial to calculate `next_cache_start` here.
|
| 191 |
+
new_cache = torch.cat((k, v), dim=-1)
|
| 192 |
+
|
| 193 |
+
scores = torch.matmul(q, k.transpose(-2, -1)) / math.sqrt(self.d_k)
|
| 194 |
+
return self.forward_attention(v, scores, mask), new_cache
|
| 195 |
+
|
| 196 |
+
|
| 197 |
+
class RelPositionMultiHeadedAttention(MultiHeadedAttention):
|
| 198 |
+
"""Multi-Head Attention layer with relative position encoding.
|
| 199 |
+
Paper: https://arxiv.org/abs/1901.02860
|
| 200 |
+
Args:
|
| 201 |
+
n_head (int): The number of heads.
|
| 202 |
+
n_feat (int): The number of features.
|
| 203 |
+
dropout_rate (float): Dropout rate.
|
| 204 |
+
"""
|
| 205 |
+
|
| 206 |
+
def __init__(
|
| 207 |
+
self, n_head: int, n_feat: int, dropout_rate: float, key_bias: bool = True
|
| 208 |
+
):
|
| 209 |
+
"""Construct an RelPositionMultiHeadedAttention object."""
|
| 210 |
+
super().__init__(n_head, n_feat, dropout_rate, key_bias)
|
| 211 |
+
# linear transformation for positional encoding
|
| 212 |
+
self.linear_pos = nn.Linear(n_feat, n_feat, bias=False)
|
| 213 |
+
# these two learnable bias are used in matrix c and matrix d
|
| 214 |
+
# as described in https://arxiv.org/abs/1901.02860 Section 3.3
|
| 215 |
+
self.pos_bias_u = nn.Parameter(torch.Tensor(self.h, self.d_k))
|
| 216 |
+
self.pos_bias_v = nn.Parameter(torch.Tensor(self.h, self.d_k))
|
| 217 |
+
torch.nn.init.xavier_uniform_(self.pos_bias_u)
|
| 218 |
+
torch.nn.init.xavier_uniform_(self.pos_bias_v)
|
| 219 |
+
|
| 220 |
+
def rel_shift(self, x: torch.Tensor) -> torch.Tensor:
|
| 221 |
+
"""Compute relative positional encoding.
|
| 222 |
+
|
| 223 |
+
Args:
|
| 224 |
+
x (torch.Tensor): Input tensor (batch, head, time1, 2*time1-1).
|
| 225 |
+
time1 means the length of query vector.
|
| 226 |
+
|
| 227 |
+
Returns:
|
| 228 |
+
torch.Tensor: Output tensor.
|
| 229 |
+
|
| 230 |
+
"""
|
| 231 |
+
zero_pad = torch.zeros(
|
| 232 |
+
(x.size()[0], x.size()[1], x.size()[2], 1), device=x.device, dtype=x.dtype
|
| 233 |
+
)
|
| 234 |
+
x_padded = torch.cat([zero_pad, x], dim=-1)
|
| 235 |
+
|
| 236 |
+
x_padded = x_padded.view(x.size()[0], x.size()[1], x.size(3) + 1, x.size(2))
|
| 237 |
+
x = x_padded[:, :, 1:].view_as(x)[
|
| 238 |
+
:, :, :, : x.size(-1) // 2 + 1
|
| 239 |
+
] # only keep the positions from 0 to time2
|
| 240 |
+
return x
|
| 241 |
+
|
| 242 |
+
def forward(
|
| 243 |
+
self,
|
| 244 |
+
query: torch.Tensor,
|
| 245 |
+
key: torch.Tensor,
|
| 246 |
+
value: torch.Tensor,
|
| 247 |
+
mask: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 248 |
+
pos_emb: torch.Tensor = torch.empty(0),
|
| 249 |
+
cache: torch.Tensor = torch.zeros((0, 0, 0, 0)),
|
| 250 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 251 |
+
"""Compute 'Scaled Dot Product Attention' with rel. positional encoding.
|
| 252 |
+
Args:
|
| 253 |
+
query (torch.Tensor): Query tensor (#batch, time1, size).
|
| 254 |
+
key (torch.Tensor): Key tensor (#batch, time2, size).
|
| 255 |
+
value (torch.Tensor): Value tensor (#batch, time2, size).
|
| 256 |
+
mask (torch.Tensor): Mask tensor (#batch, 1, time2) or
|
| 257 |
+
(#batch, time1, time2), (0, 0, 0) means fake mask.
|
| 258 |
+
pos_emb (torch.Tensor): Positional embedding tensor
|
| 259 |
+
(#batch, time2, size).
|
| 260 |
+
cache (torch.Tensor): Cache tensor (1, head, cache_t, d_k * 2),
|
| 261 |
+
where `cache_t == chunk_size * num_decoding_left_chunks`
|
| 262 |
+
and `head * d_k == size`
|
| 263 |
+
Returns:
|
| 264 |
+
torch.Tensor: Output tensor (#batch, time1, d_model).
|
| 265 |
+
torch.Tensor: Cache tensor (1, head, cache_t + time1, d_k * 2)
|
| 266 |
+
where `cache_t == chunk_size * num_decoding_left_chunks`
|
| 267 |
+
and `head * d_k == size`
|
| 268 |
+
"""
|
| 269 |
+
q, k, v = self.forward_qkv(query, key, value)
|
| 270 |
+
q = q.transpose(1, 2) # (batch, time1, head, d_k)
|
| 271 |
+
|
| 272 |
+
# NOTE(xcsong):
|
| 273 |
+
# when export onnx model, for 1st chunk, we feed
|
| 274 |
+
# cache(1, head, 0, d_k * 2) (16/-1, -1/-1, 16/0 mode)
|
| 275 |
+
# or cache(1, head, real_cache_t, d_k * 2) (16/4 mode).
|
| 276 |
+
# In all modes, `if cache.size(0) > 0` will alwayse be `True`
|
| 277 |
+
# and we will always do splitting and
|
| 278 |
+
# concatnation(this will simplify onnx export). Note that
|
| 279 |
+
# it's OK to concat & split zero-shaped tensors(see code below).
|
| 280 |
+
# when export jit model, for 1st chunk, we always feed
|
| 281 |
+
# cache(0, 0, 0, 0) since jit supports dynamic if-branch.
|
| 282 |
+
# >>> a = torch.ones((1, 2, 0, 4))
|
| 283 |
+
# >>> b = torch.ones((1, 2, 3, 4))
|
| 284 |
+
# >>> c = torch.cat((a, b), dim=2)
|
| 285 |
+
# >>> torch.equal(b, c) # True
|
| 286 |
+
# >>> d = torch.split(a, 2, dim=-1)
|
| 287 |
+
# >>> torch.equal(d[0], d[1]) # True
|
| 288 |
+
if cache.size(0) > 0:
|
| 289 |
+
key_cache, value_cache = torch.split(cache, cache.size(-1) // 2, dim=-1)
|
| 290 |
+
k = torch.cat([key_cache, k], dim=2)
|
| 291 |
+
v = torch.cat([value_cache, v], dim=2)
|
| 292 |
+
# NOTE(xcsong): We do cache slicing in encoder.forward_chunk, since it's
|
| 293 |
+
# non-trivial to calculate `next_cache_start` here.
|
| 294 |
+
new_cache = torch.cat((k, v), dim=-1)
|
| 295 |
+
|
| 296 |
+
n_batch_pos = pos_emb.size(0)
|
| 297 |
+
p = self.linear_pos(pos_emb).view(n_batch_pos, -1, self.h, self.d_k)
|
| 298 |
+
p = p.transpose(1, 2) # (batch, head, time1, d_k)
|
| 299 |
+
|
| 300 |
+
# (batch, head, time1, d_k)
|
| 301 |
+
q_with_bias_u = (q + self.pos_bias_u).transpose(1, 2)
|
| 302 |
+
# (batch, head, time1, d_k)
|
| 303 |
+
q_with_bias_v = (q + self.pos_bias_v).transpose(1, 2)
|
| 304 |
+
|
| 305 |
+
# compute attention score
|
| 306 |
+
# first compute matrix a and matrix c
|
| 307 |
+
# as described in https://arxiv.org/abs/1901.02860 Section 3.3
|
| 308 |
+
# (batch, head, time1, time2)
|
| 309 |
+
matrix_ac = torch.matmul(q_with_bias_u, k.transpose(-2, -1))
|
| 310 |
+
|
| 311 |
+
# compute matrix b and matrix d
|
| 312 |
+
# (batch, head, time1, time2)
|
| 313 |
+
matrix_bd = torch.matmul(q_with_bias_v, p.transpose(-2, -1))
|
| 314 |
+
# NOTE(Xiang Lyu): Keep rel_shift since espnet rel_pos_emb is used
|
| 315 |
+
if matrix_ac.shape != matrix_bd.shape:
|
| 316 |
+
matrix_bd = self.rel_shift(matrix_bd)
|
| 317 |
+
|
| 318 |
+
scores = (matrix_ac + matrix_bd) / math.sqrt(
|
| 319 |
+
self.d_k
|
| 320 |
+
) # (batch, head, time1, time2)
|
| 321 |
+
|
| 322 |
+
return self.forward_attention(v, scores, mask), new_cache
|
cosyvoice/transformer/convolution.py
ADDED
|
@@ -0,0 +1,147 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2020 Mobvoi Inc. (authors: Binbin Zhang, Di Wu)
|
| 2 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 16 |
+
"""ConvolutionModule definition."""
|
| 17 |
+
|
| 18 |
+
from typing import Tuple
|
| 19 |
+
|
| 20 |
+
import torch
|
| 21 |
+
from torch import nn
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
class ConvolutionModule(nn.Module):
|
| 25 |
+
"""ConvolutionModule in Conformer model."""
|
| 26 |
+
|
| 27 |
+
def __init__(
|
| 28 |
+
self,
|
| 29 |
+
channels: int,
|
| 30 |
+
kernel_size: int = 15,
|
| 31 |
+
activation: nn.Module = nn.ReLU(),
|
| 32 |
+
norm: str = "batch_norm",
|
| 33 |
+
causal: bool = False,
|
| 34 |
+
bias: bool = True,
|
| 35 |
+
):
|
| 36 |
+
"""Construct an ConvolutionModule object.
|
| 37 |
+
Args:
|
| 38 |
+
channels (int): The number of channels of conv layers.
|
| 39 |
+
kernel_size (int): Kernel size of conv layers.
|
| 40 |
+
causal (int): Whether use causal convolution or not
|
| 41 |
+
"""
|
| 42 |
+
super().__init__()
|
| 43 |
+
|
| 44 |
+
self.pointwise_conv1 = nn.Conv1d(
|
| 45 |
+
channels,
|
| 46 |
+
2 * channels,
|
| 47 |
+
kernel_size=1,
|
| 48 |
+
stride=1,
|
| 49 |
+
padding=0,
|
| 50 |
+
bias=bias,
|
| 51 |
+
)
|
| 52 |
+
# self.lorder is used to distinguish if it's a causal convolution,
|
| 53 |
+
# if self.lorder > 0: it's a causal convolution, the input will be
|
| 54 |
+
# padded with self.lorder frames on the left in forward.
|
| 55 |
+
# else: it's a symmetrical convolution
|
| 56 |
+
if causal:
|
| 57 |
+
padding = 0
|
| 58 |
+
self.lorder = kernel_size - 1
|
| 59 |
+
else:
|
| 60 |
+
# kernel_size should be an odd number for none causal convolution
|
| 61 |
+
assert (kernel_size - 1) % 2 == 0
|
| 62 |
+
padding = (kernel_size - 1) // 2
|
| 63 |
+
self.lorder = 0
|
| 64 |
+
self.depthwise_conv = nn.Conv1d(
|
| 65 |
+
channels,
|
| 66 |
+
channels,
|
| 67 |
+
kernel_size,
|
| 68 |
+
stride=1,
|
| 69 |
+
padding=padding,
|
| 70 |
+
groups=channels,
|
| 71 |
+
bias=bias,
|
| 72 |
+
)
|
| 73 |
+
|
| 74 |
+
assert norm in ["batch_norm", "layer_norm"]
|
| 75 |
+
if norm == "batch_norm":
|
| 76 |
+
self.use_layer_norm = False
|
| 77 |
+
self.norm = nn.BatchNorm1d(channels)
|
| 78 |
+
else:
|
| 79 |
+
self.use_layer_norm = True
|
| 80 |
+
self.norm = nn.LayerNorm(channels)
|
| 81 |
+
|
| 82 |
+
self.pointwise_conv2 = nn.Conv1d(
|
| 83 |
+
channels,
|
| 84 |
+
channels,
|
| 85 |
+
kernel_size=1,
|
| 86 |
+
stride=1,
|
| 87 |
+
padding=0,
|
| 88 |
+
bias=bias,
|
| 89 |
+
)
|
| 90 |
+
self.activation = activation
|
| 91 |
+
|
| 92 |
+
def forward(
|
| 93 |
+
self,
|
| 94 |
+
x: torch.Tensor,
|
| 95 |
+
mask_pad: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 96 |
+
cache: torch.Tensor = torch.zeros((0, 0, 0)),
|
| 97 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 98 |
+
"""Compute convolution module.
|
| 99 |
+
Args:
|
| 100 |
+
x (torch.Tensor): Input tensor (#batch, time, channels).
|
| 101 |
+
mask_pad (torch.Tensor): used for batch padding (#batch, 1, time),
|
| 102 |
+
(0, 0, 0) means fake mask.
|
| 103 |
+
cache (torch.Tensor): left context cache, it is only
|
| 104 |
+
used in causal convolution (#batch, channels, cache_t),
|
| 105 |
+
(0, 0, 0) meas fake cache.
|
| 106 |
+
Returns:
|
| 107 |
+
torch.Tensor: Output tensor (#batch, time, channels).
|
| 108 |
+
"""
|
| 109 |
+
# exchange the temporal dimension and the feature dimension
|
| 110 |
+
x = x.transpose(1, 2) # (#batch, channels, time)
|
| 111 |
+
|
| 112 |
+
# mask batch padding
|
| 113 |
+
if mask_pad.size(2) > 0: # time > 0
|
| 114 |
+
x.masked_fill_(~mask_pad, 0.0)
|
| 115 |
+
|
| 116 |
+
if self.lorder > 0:
|
| 117 |
+
if cache.size(2) == 0: # cache_t == 0
|
| 118 |
+
x = nn.functional.pad(x, (self.lorder, 0), "constant", 0.0)
|
| 119 |
+
else:
|
| 120 |
+
assert cache.size(0) == x.size(0) # equal batch
|
| 121 |
+
assert cache.size(1) == x.size(1) # equal channel
|
| 122 |
+
x = torch.cat((cache, x), dim=2)
|
| 123 |
+
assert x.size(2) > self.lorder
|
| 124 |
+
new_cache = x[:, :, -self.lorder :]
|
| 125 |
+
else:
|
| 126 |
+
# It's better we just return None if no cache is required,
|
| 127 |
+
# However, for JIT export, here we just fake one tensor instead of
|
| 128 |
+
# None.
|
| 129 |
+
new_cache = torch.zeros((0, 0, 0), dtype=x.dtype, device=x.device)
|
| 130 |
+
|
| 131 |
+
# GLU mechanism
|
| 132 |
+
x = self.pointwise_conv1(x) # (batch, 2*channel, dim)
|
| 133 |
+
x = nn.functional.glu(x, dim=1) # (batch, channel, dim)
|
| 134 |
+
|
| 135 |
+
# 1D Depthwise Conv
|
| 136 |
+
x = self.depthwise_conv(x)
|
| 137 |
+
if self.use_layer_norm:
|
| 138 |
+
x = x.transpose(1, 2)
|
| 139 |
+
x = self.activation(self.norm(x))
|
| 140 |
+
if self.use_layer_norm:
|
| 141 |
+
x = x.transpose(1, 2)
|
| 142 |
+
x = self.pointwise_conv2(x)
|
| 143 |
+
# mask batch padding
|
| 144 |
+
if mask_pad.size(2) > 0: # time > 0
|
| 145 |
+
x.masked_fill_(~mask_pad, 0.0)
|
| 146 |
+
|
| 147 |
+
return x.transpose(1, 2), new_cache
|
cosyvoice/transformer/decoder.py
ADDED
|
@@ -0,0 +1,418 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2021 Mobvoi Inc. (authors: Binbin Zhang, Di Wu)
|
| 2 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 16 |
+
"""Decoder definition."""
|
| 17 |
+
from typing import Tuple, List, Optional
|
| 18 |
+
|
| 19 |
+
import torch
|
| 20 |
+
import torch.utils.checkpoint as ckpt
|
| 21 |
+
import logging
|
| 22 |
+
|
| 23 |
+
from cosyvoice.transformer.decoder_layer import DecoderLayer
|
| 24 |
+
from cosyvoice.transformer.positionwise_feed_forward import (
|
| 25 |
+
PositionwiseFeedForward,
|
| 26 |
+
)
|
| 27 |
+
from cosyvoice.utils.class_utils import (
|
| 28 |
+
COSYVOICE_EMB_CLASSES,
|
| 29 |
+
COSYVOICE_ATTENTION_CLASSES,
|
| 30 |
+
COSYVOICE_ACTIVATION_CLASSES,
|
| 31 |
+
)
|
| 32 |
+
from cosyvoice.utils.mask import subsequent_mask, make_pad_mask
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
class TransformerDecoder(torch.nn.Module):
|
| 36 |
+
"""Base class of Transfomer decoder module.
|
| 37 |
+
Args:
|
| 38 |
+
vocab_size: output dim
|
| 39 |
+
encoder_output_size: dimension of attention
|
| 40 |
+
attention_heads: the number of heads of multi head attention
|
| 41 |
+
linear_units: the hidden units number of position-wise feedforward
|
| 42 |
+
num_blocks: the number of decoder blocks
|
| 43 |
+
dropout_rate: dropout rate
|
| 44 |
+
self_attention_dropout_rate: dropout rate for attention
|
| 45 |
+
input_layer: input layer type
|
| 46 |
+
use_output_layer: whether to use output layer
|
| 47 |
+
pos_enc_class: PositionalEncoding or ScaledPositionalEncoding
|
| 48 |
+
normalize_before:
|
| 49 |
+
True: use layer_norm before each sub-block of a layer.
|
| 50 |
+
False: use layer_norm after each sub-block of a layer.
|
| 51 |
+
src_attention: if false, encoder-decoder cross attention is not
|
| 52 |
+
applied, such as CIF model
|
| 53 |
+
key_bias: whether use bias in attention.linear_k, False for whisper models.
|
| 54 |
+
gradient_checkpointing: rerunning a forward-pass segment for each
|
| 55 |
+
checkpointed segment during backward.
|
| 56 |
+
tie_word_embedding: Tie or clone module weights depending of whether we are
|
| 57 |
+
using TorchScript or not
|
| 58 |
+
"""
|
| 59 |
+
|
| 60 |
+
def __init__(
|
| 61 |
+
self,
|
| 62 |
+
vocab_size: int,
|
| 63 |
+
encoder_output_size: int,
|
| 64 |
+
attention_heads: int = 4,
|
| 65 |
+
linear_units: int = 2048,
|
| 66 |
+
num_blocks: int = 6,
|
| 67 |
+
dropout_rate: float = 0.1,
|
| 68 |
+
positional_dropout_rate: float = 0.1,
|
| 69 |
+
self_attention_dropout_rate: float = 0.0,
|
| 70 |
+
src_attention_dropout_rate: float = 0.0,
|
| 71 |
+
input_layer: str = "embed",
|
| 72 |
+
use_output_layer: bool = True,
|
| 73 |
+
normalize_before: bool = True,
|
| 74 |
+
src_attention: bool = True,
|
| 75 |
+
key_bias: bool = True,
|
| 76 |
+
activation_type: str = "relu",
|
| 77 |
+
gradient_checkpointing: bool = False,
|
| 78 |
+
tie_word_embedding: bool = False,
|
| 79 |
+
):
|
| 80 |
+
super().__init__()
|
| 81 |
+
attention_dim = encoder_output_size
|
| 82 |
+
activation = COSYVOICE_ACTIVATION_CLASSES[activation_type]()
|
| 83 |
+
|
| 84 |
+
self.embed = torch.nn.Sequential(
|
| 85 |
+
(
|
| 86 |
+
torch.nn.Identity()
|
| 87 |
+
if input_layer == "no_pos"
|
| 88 |
+
else torch.nn.Embedding(vocab_size, attention_dim)
|
| 89 |
+
),
|
| 90 |
+
COSYVOICE_EMB_CLASSES[input_layer](attention_dim, positional_dropout_rate),
|
| 91 |
+
)
|
| 92 |
+
|
| 93 |
+
self.normalize_before = normalize_before
|
| 94 |
+
self.after_norm = torch.nn.LayerNorm(attention_dim, eps=1e-5)
|
| 95 |
+
self.use_output_layer = use_output_layer
|
| 96 |
+
if use_output_layer:
|
| 97 |
+
self.output_layer = torch.nn.Linear(attention_dim, vocab_size)
|
| 98 |
+
else:
|
| 99 |
+
self.output_layer = torch.nn.Identity()
|
| 100 |
+
self.num_blocks = num_blocks
|
| 101 |
+
self.decoders = torch.nn.ModuleList(
|
| 102 |
+
[
|
| 103 |
+
DecoderLayer(
|
| 104 |
+
attention_dim,
|
| 105 |
+
COSYVOICE_ATTENTION_CLASSES["selfattn"](
|
| 106 |
+
attention_heads,
|
| 107 |
+
attention_dim,
|
| 108 |
+
self_attention_dropout_rate,
|
| 109 |
+
key_bias,
|
| 110 |
+
),
|
| 111 |
+
(
|
| 112 |
+
COSYVOICE_ATTENTION_CLASSES["selfattn"](
|
| 113 |
+
attention_heads,
|
| 114 |
+
attention_dim,
|
| 115 |
+
src_attention_dropout_rate,
|
| 116 |
+
key_bias,
|
| 117 |
+
)
|
| 118 |
+
if src_attention
|
| 119 |
+
else None
|
| 120 |
+
),
|
| 121 |
+
PositionwiseFeedForward(
|
| 122 |
+
attention_dim, linear_units, dropout_rate, activation
|
| 123 |
+
),
|
| 124 |
+
dropout_rate,
|
| 125 |
+
normalize_before,
|
| 126 |
+
)
|
| 127 |
+
for _ in range(self.num_blocks)
|
| 128 |
+
]
|
| 129 |
+
)
|
| 130 |
+
|
| 131 |
+
self.gradient_checkpointing = gradient_checkpointing
|
| 132 |
+
self.tie_word_embedding = tie_word_embedding
|
| 133 |
+
|
| 134 |
+
def forward(
|
| 135 |
+
self,
|
| 136 |
+
memory: torch.Tensor,
|
| 137 |
+
memory_mask: torch.Tensor,
|
| 138 |
+
ys_in_pad: torch.Tensor,
|
| 139 |
+
ys_in_lens: torch.Tensor,
|
| 140 |
+
r_ys_in_pad: torch.Tensor = torch.empty(0),
|
| 141 |
+
reverse_weight: float = 0.0,
|
| 142 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 143 |
+
"""Forward decoder.
|
| 144 |
+
Args:
|
| 145 |
+
memory: encoded memory, float32 (batch, maxlen_in, feat)
|
| 146 |
+
memory_mask: encoder memory mask, (batch, 1, maxlen_in)
|
| 147 |
+
ys_in_pad: padded input token ids, int64 (batch, maxlen_out)
|
| 148 |
+
ys_in_lens: input lengths of this batch (batch)
|
| 149 |
+
r_ys_in_pad: not used in transformer decoder, in order to unify api
|
| 150 |
+
with bidirectional decoder
|
| 151 |
+
reverse_weight: not used in transformer decoder, in order to unify
|
| 152 |
+
api with bidirectional decode
|
| 153 |
+
Returns:
|
| 154 |
+
(tuple): tuple containing:
|
| 155 |
+
x: decoded token score before softmax (batch, maxlen_out,
|
| 156 |
+
vocab_size) if use_output_layer is True,
|
| 157 |
+
torch.tensor(0.0), in order to unify api with bidirectional decoder
|
| 158 |
+
olens: (batch, )
|
| 159 |
+
NOTE(xcsong):
|
| 160 |
+
We pass the `__call__` method of the modules instead of `forward` to the
|
| 161 |
+
checkpointing API because `__call__` attaches all the hooks of the module.
|
| 162 |
+
https://discuss.pytorch.org/t/any-different-between-model-input-and-model-forward-input/3690/2
|
| 163 |
+
"""
|
| 164 |
+
tgt = ys_in_pad
|
| 165 |
+
maxlen = tgt.size(1)
|
| 166 |
+
# tgt_mask: (B, 1, L)
|
| 167 |
+
tgt_mask = ~make_pad_mask(ys_in_lens, maxlen).unsqueeze(1)
|
| 168 |
+
tgt_mask = tgt_mask.to(tgt.device)
|
| 169 |
+
# m: (1, L, L)
|
| 170 |
+
m = subsequent_mask(tgt_mask.size(-1), device=tgt_mask.device).unsqueeze(0)
|
| 171 |
+
# tgt_mask: (B, L, L)
|
| 172 |
+
tgt_mask = tgt_mask & m
|
| 173 |
+
x, _ = self.embed(tgt)
|
| 174 |
+
if self.gradient_checkpointing and self.training:
|
| 175 |
+
x = self.forward_layers_checkpointed(x, tgt_mask, memory, memory_mask)
|
| 176 |
+
else:
|
| 177 |
+
x = self.forward_layers(x, tgt_mask, memory, memory_mask)
|
| 178 |
+
if self.normalize_before:
|
| 179 |
+
x = self.after_norm(x)
|
| 180 |
+
if self.use_output_layer:
|
| 181 |
+
x = self.output_layer(x)
|
| 182 |
+
olens = tgt_mask.sum(1)
|
| 183 |
+
return x, torch.tensor(0.0), olens
|
| 184 |
+
|
| 185 |
+
def forward_layers(
|
| 186 |
+
self,
|
| 187 |
+
x: torch.Tensor,
|
| 188 |
+
tgt_mask: torch.Tensor,
|
| 189 |
+
memory: torch.Tensor,
|
| 190 |
+
memory_mask: torch.Tensor,
|
| 191 |
+
) -> torch.Tensor:
|
| 192 |
+
for layer in self.decoders:
|
| 193 |
+
x, tgt_mask, memory, memory_mask = layer(x, tgt_mask, memory, memory_mask)
|
| 194 |
+
return x
|
| 195 |
+
|
| 196 |
+
@torch.jit.unused
|
| 197 |
+
def forward_layers_checkpointed(
|
| 198 |
+
self,
|
| 199 |
+
x: torch.Tensor,
|
| 200 |
+
tgt_mask: torch.Tensor,
|
| 201 |
+
memory: torch.Tensor,
|
| 202 |
+
memory_mask: torch.Tensor,
|
| 203 |
+
) -> torch.Tensor:
|
| 204 |
+
for layer in self.decoders:
|
| 205 |
+
x, tgt_mask, memory, memory_mask = ckpt.checkpoint(
|
| 206 |
+
layer.__call__, x, tgt_mask, memory, memory_mask
|
| 207 |
+
)
|
| 208 |
+
return x
|
| 209 |
+
|
| 210 |
+
def forward_one_step(
|
| 211 |
+
self,
|
| 212 |
+
memory: torch.Tensor,
|
| 213 |
+
memory_mask: torch.Tensor,
|
| 214 |
+
tgt: torch.Tensor,
|
| 215 |
+
tgt_mask: torch.Tensor,
|
| 216 |
+
cache: Optional[List[torch.Tensor]] = None,
|
| 217 |
+
) -> Tuple[torch.Tensor, List[torch.Tensor]]:
|
| 218 |
+
"""Forward one step.
|
| 219 |
+
This is only used for decoding.
|
| 220 |
+
Args:
|
| 221 |
+
memory: encoded memory, float32 (batch, maxlen_in, feat)
|
| 222 |
+
memory_mask: encoded memory mask, (batch, 1, maxlen_in)
|
| 223 |
+
tgt: input token ids, int64 (batch, maxlen_out)
|
| 224 |
+
tgt_mask: input token mask, (batch, maxlen_out)
|
| 225 |
+
dtype=torch.uint8 in PyTorch 1.2-
|
| 226 |
+
dtype=torch.bool in PyTorch 1.2+ (include 1.2)
|
| 227 |
+
cache: cached output list of (batch, max_time_out-1, size)
|
| 228 |
+
Returns:
|
| 229 |
+
y, cache: NN output value and cache per `self.decoders`.
|
| 230 |
+
y.shape` is (batch, maxlen_out, token)
|
| 231 |
+
"""
|
| 232 |
+
x, _ = self.embed(tgt)
|
| 233 |
+
new_cache = []
|
| 234 |
+
for i, decoder in enumerate(self.decoders):
|
| 235 |
+
if cache is None:
|
| 236 |
+
c = None
|
| 237 |
+
else:
|
| 238 |
+
c = cache[i]
|
| 239 |
+
x, tgt_mask, memory, memory_mask = decoder(
|
| 240 |
+
x, tgt_mask, memory, memory_mask, cache=c
|
| 241 |
+
)
|
| 242 |
+
new_cache.append(x)
|
| 243 |
+
if self.normalize_before:
|
| 244 |
+
y = self.after_norm(x[:, -1])
|
| 245 |
+
else:
|
| 246 |
+
y = x[:, -1]
|
| 247 |
+
if self.use_output_layer:
|
| 248 |
+
y = torch.log_softmax(self.output_layer(y), dim=-1)
|
| 249 |
+
return y, new_cache
|
| 250 |
+
|
| 251 |
+
def tie_or_clone_weights(self, jit_mode: bool = True):
|
| 252 |
+
"""Tie or clone module weights (between word_emb and output_layer)
|
| 253 |
+
depending of whether we are using TorchScript or not"""
|
| 254 |
+
if not self.use_output_layer:
|
| 255 |
+
return
|
| 256 |
+
if jit_mode:
|
| 257 |
+
logging.info("clone emb.weight to output.weight")
|
| 258 |
+
self.output_layer.weight = torch.nn.Parameter(self.embed[0].weight.clone())
|
| 259 |
+
else:
|
| 260 |
+
logging.info("tie emb.weight with output.weight")
|
| 261 |
+
self.output_layer.weight = self.embed[0].weight
|
| 262 |
+
|
| 263 |
+
if getattr(self.output_layer, "bias", None) is not None:
|
| 264 |
+
self.output_layer.bias.data = torch.nn.functional.pad(
|
| 265 |
+
self.output_layer.bias.data,
|
| 266 |
+
(
|
| 267 |
+
0,
|
| 268 |
+
self.output_layer.weight.shape[0] - self.output_layer.bias.shape[0],
|
| 269 |
+
),
|
| 270 |
+
"constant",
|
| 271 |
+
0,
|
| 272 |
+
)
|
| 273 |
+
|
| 274 |
+
|
| 275 |
+
class BiTransformerDecoder(torch.nn.Module):
|
| 276 |
+
"""Base class of Transfomer decoder module.
|
| 277 |
+
Args:
|
| 278 |
+
vocab_size: output dim
|
| 279 |
+
encoder_output_size: dimension of attention
|
| 280 |
+
attention_heads: the number of heads of multi head attention
|
| 281 |
+
linear_units: the hidden units number of position-wise feedforward
|
| 282 |
+
num_blocks: the number of decoder blocks
|
| 283 |
+
r_num_blocks: the number of right to left decoder blocks
|
| 284 |
+
dropout_rate: dropout rate
|
| 285 |
+
self_attention_dropout_rate: dropout rate for attention
|
| 286 |
+
input_layer: input layer type
|
| 287 |
+
use_output_layer: whether to use output layer
|
| 288 |
+
pos_enc_class: PositionalEncoding or ScaledPositionalEncoding
|
| 289 |
+
normalize_before:
|
| 290 |
+
True: use layer_norm before each sub-block of a layer.
|
| 291 |
+
False: use layer_norm after each sub-block of a layer.
|
| 292 |
+
key_bias: whether use bias in attention.linear_k, False for whisper models.
|
| 293 |
+
"""
|
| 294 |
+
|
| 295 |
+
def __init__(
|
| 296 |
+
self,
|
| 297 |
+
vocab_size: int,
|
| 298 |
+
encoder_output_size: int,
|
| 299 |
+
attention_heads: int = 4,
|
| 300 |
+
linear_units: int = 2048,
|
| 301 |
+
num_blocks: int = 6,
|
| 302 |
+
r_num_blocks: int = 0,
|
| 303 |
+
dropout_rate: float = 0.1,
|
| 304 |
+
positional_dropout_rate: float = 0.1,
|
| 305 |
+
self_attention_dropout_rate: float = 0.0,
|
| 306 |
+
src_attention_dropout_rate: float = 0.0,
|
| 307 |
+
input_layer: str = "embed",
|
| 308 |
+
use_output_layer: bool = True,
|
| 309 |
+
normalize_before: bool = True,
|
| 310 |
+
key_bias: bool = True,
|
| 311 |
+
gradient_checkpointing: bool = False,
|
| 312 |
+
tie_word_embedding: bool = False,
|
| 313 |
+
):
|
| 314 |
+
|
| 315 |
+
super().__init__()
|
| 316 |
+
self.tie_word_embedding = tie_word_embedding
|
| 317 |
+
self.left_decoder = TransformerDecoder(
|
| 318 |
+
vocab_size,
|
| 319 |
+
encoder_output_size,
|
| 320 |
+
attention_heads,
|
| 321 |
+
linear_units,
|
| 322 |
+
num_blocks,
|
| 323 |
+
dropout_rate,
|
| 324 |
+
positional_dropout_rate,
|
| 325 |
+
self_attention_dropout_rate,
|
| 326 |
+
src_attention_dropout_rate,
|
| 327 |
+
input_layer,
|
| 328 |
+
use_output_layer,
|
| 329 |
+
normalize_before,
|
| 330 |
+
key_bias=key_bias,
|
| 331 |
+
gradient_checkpointing=gradient_checkpointing,
|
| 332 |
+
tie_word_embedding=tie_word_embedding,
|
| 333 |
+
)
|
| 334 |
+
|
| 335 |
+
self.right_decoder = TransformerDecoder(
|
| 336 |
+
vocab_size,
|
| 337 |
+
encoder_output_size,
|
| 338 |
+
attention_heads,
|
| 339 |
+
linear_units,
|
| 340 |
+
r_num_blocks,
|
| 341 |
+
dropout_rate,
|
| 342 |
+
positional_dropout_rate,
|
| 343 |
+
self_attention_dropout_rate,
|
| 344 |
+
src_attention_dropout_rate,
|
| 345 |
+
input_layer,
|
| 346 |
+
use_output_layer,
|
| 347 |
+
normalize_before,
|
| 348 |
+
key_bias=key_bias,
|
| 349 |
+
gradient_checkpointing=gradient_checkpointing,
|
| 350 |
+
tie_word_embedding=tie_word_embedding,
|
| 351 |
+
)
|
| 352 |
+
|
| 353 |
+
def forward(
|
| 354 |
+
self,
|
| 355 |
+
memory: torch.Tensor,
|
| 356 |
+
memory_mask: torch.Tensor,
|
| 357 |
+
ys_in_pad: torch.Tensor,
|
| 358 |
+
ys_in_lens: torch.Tensor,
|
| 359 |
+
r_ys_in_pad: torch.Tensor,
|
| 360 |
+
reverse_weight: float = 0.0,
|
| 361 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 362 |
+
"""Forward decoder.
|
| 363 |
+
Args:
|
| 364 |
+
memory: encoded memory, float32 (batch, maxlen_in, feat)
|
| 365 |
+
memory_mask: encoder memory mask, (batch, 1, maxlen_in)
|
| 366 |
+
ys_in_pad: padded input token ids, int64 (batch, maxlen_out)
|
| 367 |
+
ys_in_lens: input lengths of this batch (batch)
|
| 368 |
+
r_ys_in_pad: padded input token ids, int64 (batch, maxlen_out),
|
| 369 |
+
used for right to left decoder
|
| 370 |
+
reverse_weight: used for right to left decoder
|
| 371 |
+
Returns:
|
| 372 |
+
(tuple): tuple containing:
|
| 373 |
+
x: decoded token score before softmax (batch, maxlen_out,
|
| 374 |
+
vocab_size) if use_output_layer is True,
|
| 375 |
+
r_x: x: decoded token score (right to left decoder)
|
| 376 |
+
before softmax (batch, maxlen_out, vocab_size)
|
| 377 |
+
if use_output_layer is True,
|
| 378 |
+
olens: (batch, )
|
| 379 |
+
"""
|
| 380 |
+
l_x, _, olens = self.left_decoder(memory, memory_mask, ys_in_pad, ys_in_lens)
|
| 381 |
+
r_x = torch.tensor(0.0)
|
| 382 |
+
if reverse_weight > 0.0:
|
| 383 |
+
r_x, _, olens = self.right_decoder(
|
| 384 |
+
memory, memory_mask, r_ys_in_pad, ys_in_lens
|
| 385 |
+
)
|
| 386 |
+
return l_x, r_x, olens
|
| 387 |
+
|
| 388 |
+
def forward_one_step(
|
| 389 |
+
self,
|
| 390 |
+
memory: torch.Tensor,
|
| 391 |
+
memory_mask: torch.Tensor,
|
| 392 |
+
tgt: torch.Tensor,
|
| 393 |
+
tgt_mask: torch.Tensor,
|
| 394 |
+
cache: Optional[List[torch.Tensor]] = None,
|
| 395 |
+
) -> Tuple[torch.Tensor, List[torch.Tensor]]:
|
| 396 |
+
"""Forward one step.
|
| 397 |
+
This is only used for decoding.
|
| 398 |
+
Args:
|
| 399 |
+
memory: encoded memory, float32 (batch, maxlen_in, feat)
|
| 400 |
+
memory_mask: encoded memory mask, (batch, 1, maxlen_in)
|
| 401 |
+
tgt: input token ids, int64 (batch, maxlen_out)
|
| 402 |
+
tgt_mask: input token mask, (batch, maxlen_out)
|
| 403 |
+
dtype=torch.uint8 in PyTorch 1.2-
|
| 404 |
+
dtype=torch.bool in PyTorch 1.2+ (include 1.2)
|
| 405 |
+
cache: cached output list of (batch, max_time_out-1, size)
|
| 406 |
+
Returns:
|
| 407 |
+
y, cache: NN output value and cache per `self.decoders`.
|
| 408 |
+
y.shape` is (batch, maxlen_out, token)
|
| 409 |
+
"""
|
| 410 |
+
return self.left_decoder.forward_one_step(
|
| 411 |
+
memory, memory_mask, tgt, tgt_mask, cache
|
| 412 |
+
)
|
| 413 |
+
|
| 414 |
+
def tie_or_clone_weights(self, jit_mode: bool = True):
|
| 415 |
+
"""Tie or clone module weights (between word_emb and output_layer)
|
| 416 |
+
depending of whether we are using TorchScript or not"""
|
| 417 |
+
self.left_decoder.tie_or_clone_weights(jit_mode)
|
| 418 |
+
self.right_decoder.tie_or_clone_weights(jit_mode)
|
cosyvoice/transformer/decoder_layer.py
ADDED
|
@@ -0,0 +1,132 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2019 Shigeki Karita
|
| 2 |
+
# 2020 Mobvoi Inc (Binbin Zhang)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
"""Decoder self-attention layer definition."""
|
| 16 |
+
from typing import Optional, Tuple
|
| 17 |
+
|
| 18 |
+
import torch
|
| 19 |
+
from torch import nn
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
class DecoderLayer(nn.Module):
|
| 23 |
+
"""Single decoder layer module.
|
| 24 |
+
|
| 25 |
+
Args:
|
| 26 |
+
size (int): Input dimension.
|
| 27 |
+
self_attn (torch.nn.Module): Self-attention module instance.
|
| 28 |
+
`MultiHeadedAttention` instance can be used as the argument.
|
| 29 |
+
src_attn (torch.nn.Module): Inter-attention module instance.
|
| 30 |
+
`MultiHeadedAttention` instance can be used as the argument.
|
| 31 |
+
If `None` is passed, Inter-attention is not used, such as
|
| 32 |
+
CIF, GPT, and other decoder only model.
|
| 33 |
+
feed_forward (torch.nn.Module): Feed-forward module instance.
|
| 34 |
+
`PositionwiseFeedForward` instance can be used as the argument.
|
| 35 |
+
dropout_rate (float): Dropout rate.
|
| 36 |
+
normalize_before (bool):
|
| 37 |
+
True: use layer_norm before each sub-block.
|
| 38 |
+
False: to use layer_norm after each sub-block.
|
| 39 |
+
"""
|
| 40 |
+
|
| 41 |
+
def __init__(
|
| 42 |
+
self,
|
| 43 |
+
size: int,
|
| 44 |
+
self_attn: nn.Module,
|
| 45 |
+
src_attn: Optional[nn.Module],
|
| 46 |
+
feed_forward: nn.Module,
|
| 47 |
+
dropout_rate: float,
|
| 48 |
+
normalize_before: bool = True,
|
| 49 |
+
):
|
| 50 |
+
"""Construct an DecoderLayer object."""
|
| 51 |
+
super().__init__()
|
| 52 |
+
self.size = size
|
| 53 |
+
self.self_attn = self_attn
|
| 54 |
+
self.src_attn = src_attn
|
| 55 |
+
self.feed_forward = feed_forward
|
| 56 |
+
self.norm1 = nn.LayerNorm(size, eps=1e-5)
|
| 57 |
+
self.norm2 = nn.LayerNorm(size, eps=1e-5)
|
| 58 |
+
self.norm3 = nn.LayerNorm(size, eps=1e-5)
|
| 59 |
+
self.dropout = nn.Dropout(dropout_rate)
|
| 60 |
+
self.normalize_before = normalize_before
|
| 61 |
+
|
| 62 |
+
def forward(
|
| 63 |
+
self,
|
| 64 |
+
tgt: torch.Tensor,
|
| 65 |
+
tgt_mask: torch.Tensor,
|
| 66 |
+
memory: torch.Tensor,
|
| 67 |
+
memory_mask: torch.Tensor,
|
| 68 |
+
cache: Optional[torch.Tensor] = None,
|
| 69 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 70 |
+
"""Compute decoded features.
|
| 71 |
+
|
| 72 |
+
Args:
|
| 73 |
+
tgt (torch.Tensor): Input tensor (#batch, maxlen_out, size).
|
| 74 |
+
tgt_mask (torch.Tensor): Mask for input tensor
|
| 75 |
+
(#batch, maxlen_out).
|
| 76 |
+
memory (torch.Tensor): Encoded memory
|
| 77 |
+
(#batch, maxlen_in, size).
|
| 78 |
+
memory_mask (torch.Tensor): Encoded memory mask
|
| 79 |
+
(#batch, maxlen_in).
|
| 80 |
+
cache (torch.Tensor): cached tensors.
|
| 81 |
+
(#batch, maxlen_out - 1, size).
|
| 82 |
+
|
| 83 |
+
Returns:
|
| 84 |
+
torch.Tensor: Output tensor (#batch, maxlen_out, size).
|
| 85 |
+
torch.Tensor: Mask for output tensor (#batch, maxlen_out).
|
| 86 |
+
torch.Tensor: Encoded memory (#batch, maxlen_in, size).
|
| 87 |
+
torch.Tensor: Encoded memory mask (#batch, maxlen_in).
|
| 88 |
+
|
| 89 |
+
"""
|
| 90 |
+
residual = tgt
|
| 91 |
+
if self.normalize_before:
|
| 92 |
+
tgt = self.norm1(tgt)
|
| 93 |
+
|
| 94 |
+
if cache is None:
|
| 95 |
+
tgt_q = tgt
|
| 96 |
+
tgt_q_mask = tgt_mask
|
| 97 |
+
else:
|
| 98 |
+
# compute only the last frame query keeping dim: max_time_out -> 1
|
| 99 |
+
assert cache.shape == (
|
| 100 |
+
tgt.shape[0],
|
| 101 |
+
tgt.shape[1] - 1,
|
| 102 |
+
self.size,
|
| 103 |
+
), "{cache.shape} == {(tgt.shape[0], tgt.shape[1] - 1, self.size)}"
|
| 104 |
+
tgt_q = tgt[:, -1:, :]
|
| 105 |
+
residual = residual[:, -1:, :]
|
| 106 |
+
tgt_q_mask = tgt_mask[:, -1:, :]
|
| 107 |
+
|
| 108 |
+
x = residual + self.dropout(self.self_attn(tgt_q, tgt, tgt, tgt_q_mask)[0])
|
| 109 |
+
if not self.normalize_before:
|
| 110 |
+
x = self.norm1(x)
|
| 111 |
+
|
| 112 |
+
if self.src_attn is not None:
|
| 113 |
+
residual = x
|
| 114 |
+
if self.normalize_before:
|
| 115 |
+
x = self.norm2(x)
|
| 116 |
+
x = residual + self.dropout(
|
| 117 |
+
self.src_attn(x, memory, memory, memory_mask)[0]
|
| 118 |
+
)
|
| 119 |
+
if not self.normalize_before:
|
| 120 |
+
x = self.norm2(x)
|
| 121 |
+
|
| 122 |
+
residual = x
|
| 123 |
+
if self.normalize_before:
|
| 124 |
+
x = self.norm3(x)
|
| 125 |
+
x = residual + self.dropout(self.feed_forward(x))
|
| 126 |
+
if not self.normalize_before:
|
| 127 |
+
x = self.norm3(x)
|
| 128 |
+
|
| 129 |
+
if cache is not None:
|
| 130 |
+
x = torch.cat([cache, x], dim=1)
|
| 131 |
+
|
| 132 |
+
return x, tgt_mask, memory, memory_mask
|
cosyvoice/transformer/embedding.py
ADDED
|
@@ -0,0 +1,293 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2020 Mobvoi Inc. (authors: Binbin Zhang, Di Wu)
|
| 2 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 16 |
+
"""Positonal Encoding Module."""
|
| 17 |
+
|
| 18 |
+
import math
|
| 19 |
+
from typing import Tuple, Union
|
| 20 |
+
|
| 21 |
+
import torch
|
| 22 |
+
import torch.nn.functional as F
|
| 23 |
+
import numpy as np
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
class PositionalEncoding(torch.nn.Module):
|
| 27 |
+
"""Positional encoding.
|
| 28 |
+
|
| 29 |
+
:param int d_model: embedding dim
|
| 30 |
+
:param float dropout_rate: dropout rate
|
| 31 |
+
:param int max_len: maximum input length
|
| 32 |
+
|
| 33 |
+
PE(pos, 2i) = sin(pos/(10000^(2i/dmodel)))
|
| 34 |
+
PE(pos, 2i+1) = cos(pos/(10000^(2i/dmodel)))
|
| 35 |
+
"""
|
| 36 |
+
|
| 37 |
+
def __init__(
|
| 38 |
+
self,
|
| 39 |
+
d_model: int,
|
| 40 |
+
dropout_rate: float,
|
| 41 |
+
max_len: int = 5000,
|
| 42 |
+
reverse: bool = False,
|
| 43 |
+
):
|
| 44 |
+
"""Construct an PositionalEncoding object."""
|
| 45 |
+
super().__init__()
|
| 46 |
+
self.d_model = d_model
|
| 47 |
+
self.xscale = math.sqrt(self.d_model)
|
| 48 |
+
self.dropout = torch.nn.Dropout(p=dropout_rate)
|
| 49 |
+
self.max_len = max_len
|
| 50 |
+
|
| 51 |
+
self.pe = torch.zeros(self.max_len, self.d_model)
|
| 52 |
+
position = torch.arange(0, self.max_len, dtype=torch.float32).unsqueeze(1)
|
| 53 |
+
div_term = torch.exp(
|
| 54 |
+
torch.arange(0, self.d_model, 2, dtype=torch.float32)
|
| 55 |
+
* -(math.log(10000.0) / self.d_model)
|
| 56 |
+
)
|
| 57 |
+
self.pe[:, 0::2] = torch.sin(position * div_term)
|
| 58 |
+
self.pe[:, 1::2] = torch.cos(position * div_term)
|
| 59 |
+
self.pe = self.pe.unsqueeze(0)
|
| 60 |
+
|
| 61 |
+
def forward(
|
| 62 |
+
self, x: torch.Tensor, offset: Union[int, torch.Tensor] = 0
|
| 63 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 64 |
+
"""Add positional encoding.
|
| 65 |
+
|
| 66 |
+
Args:
|
| 67 |
+
x (torch.Tensor): Input. Its shape is (batch, time, ...)
|
| 68 |
+
offset (int, torch.tensor): position offset
|
| 69 |
+
|
| 70 |
+
Returns:
|
| 71 |
+
torch.Tensor: Encoded tensor. Its shape is (batch, time, ...)
|
| 72 |
+
torch.Tensor: for compatibility to RelPositionalEncoding
|
| 73 |
+
"""
|
| 74 |
+
|
| 75 |
+
self.pe = self.pe.to(x.device)
|
| 76 |
+
pos_emb = self.position_encoding(offset, x.size(1), False)
|
| 77 |
+
x = x * self.xscale + pos_emb
|
| 78 |
+
return self.dropout(x), self.dropout(pos_emb)
|
| 79 |
+
|
| 80 |
+
def position_encoding(
|
| 81 |
+
self, offset: Union[int, torch.Tensor], size: int, apply_dropout: bool = True
|
| 82 |
+
) -> torch.Tensor:
|
| 83 |
+
"""For getting encoding in a streaming fashion
|
| 84 |
+
|
| 85 |
+
Attention!!!!!
|
| 86 |
+
we apply dropout only once at the whole utterance level in a none
|
| 87 |
+
streaming way, but will call this function several times with
|
| 88 |
+
increasing input size in a streaming scenario, so the dropout will
|
| 89 |
+
be applied several times.
|
| 90 |
+
|
| 91 |
+
Args:
|
| 92 |
+
offset (int or torch.tensor): start offset
|
| 93 |
+
size (int): required size of position encoding
|
| 94 |
+
|
| 95 |
+
Returns:
|
| 96 |
+
torch.Tensor: Corresponding encoding
|
| 97 |
+
"""
|
| 98 |
+
# How to subscript a Union type:
|
| 99 |
+
# https://github.com/pytorch/pytorch/issues/69434
|
| 100 |
+
if isinstance(offset, int):
|
| 101 |
+
assert offset + size <= self.max_len
|
| 102 |
+
pos_emb = self.pe[:, offset : offset + size]
|
| 103 |
+
elif isinstance(offset, torch.Tensor) and offset.dim() == 0: # scalar
|
| 104 |
+
assert offset + size <= self.max_len
|
| 105 |
+
pos_emb = self.pe[:, offset : offset + size]
|
| 106 |
+
else: # for batched streaming decoding on GPU
|
| 107 |
+
assert torch.max(offset) + size <= self.max_len
|
| 108 |
+
index = offset.unsqueeze(1) + torch.arange(0, size).to(
|
| 109 |
+
offset.device
|
| 110 |
+
) # B X T
|
| 111 |
+
flag = index > 0
|
| 112 |
+
# remove negative offset
|
| 113 |
+
index = index * flag
|
| 114 |
+
pos_emb = F.embedding(index, self.pe[0]) # B X T X d_model
|
| 115 |
+
|
| 116 |
+
if apply_dropout:
|
| 117 |
+
pos_emb = self.dropout(pos_emb)
|
| 118 |
+
return pos_emb
|
| 119 |
+
|
| 120 |
+
|
| 121 |
+
class RelPositionalEncoding(PositionalEncoding):
|
| 122 |
+
"""Relative positional encoding module.
|
| 123 |
+
See : Appendix B in https://arxiv.org/abs/1901.02860
|
| 124 |
+
Args:
|
| 125 |
+
d_model (int): Embedding dimension.
|
| 126 |
+
dropout_rate (float): Dropout rate.
|
| 127 |
+
max_len (int): Maximum input length.
|
| 128 |
+
"""
|
| 129 |
+
|
| 130 |
+
def __init__(self, d_model: int, dropout_rate: float, max_len: int = 5000):
|
| 131 |
+
"""Initialize class."""
|
| 132 |
+
super().__init__(d_model, dropout_rate, max_len, reverse=True)
|
| 133 |
+
|
| 134 |
+
def forward(
|
| 135 |
+
self, x: torch.Tensor, offset: Union[int, torch.Tensor] = 0
|
| 136 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 137 |
+
"""Compute positional encoding.
|
| 138 |
+
Args:
|
| 139 |
+
x (torch.Tensor): Input tensor (batch, time, `*`).
|
| 140 |
+
Returns:
|
| 141 |
+
torch.Tensor: Encoded tensor (batch, time, `*`).
|
| 142 |
+
torch.Tensor: Positional embedding tensor (1, time, `*`).
|
| 143 |
+
"""
|
| 144 |
+
self.pe = self.pe.to(x.device)
|
| 145 |
+
x = x * self.xscale
|
| 146 |
+
pos_emb = self.position_encoding(offset, x.size(1), False)
|
| 147 |
+
return self.dropout(x), self.dropout(pos_emb)
|
| 148 |
+
|
| 149 |
+
|
| 150 |
+
class WhisperPositionalEncoding(PositionalEncoding):
|
| 151 |
+
"""Sinusoids position encoding used in openai-whisper.encoder"""
|
| 152 |
+
|
| 153 |
+
def __init__(self, d_model: int, dropout_rate: float, max_len: int = 1500):
|
| 154 |
+
super().__init__(d_model, dropout_rate, max_len)
|
| 155 |
+
self.xscale = 1.0
|
| 156 |
+
log_timescale_increment = np.log(10000) / (d_model // 2 - 1)
|
| 157 |
+
inv_timescales = torch.exp(
|
| 158 |
+
-log_timescale_increment * torch.arange(d_model // 2)
|
| 159 |
+
)
|
| 160 |
+
scaled_time = (
|
| 161 |
+
torch.arange(max_len)[:, np.newaxis] * inv_timescales[np.newaxis, :]
|
| 162 |
+
)
|
| 163 |
+
pe = torch.cat([torch.sin(scaled_time), torch.cos(scaled_time)], dim=1)
|
| 164 |
+
delattr(self, "pe")
|
| 165 |
+
self.register_buffer("pe", pe.unsqueeze(0))
|
| 166 |
+
|
| 167 |
+
|
| 168 |
+
class LearnablePositionalEncoding(PositionalEncoding):
|
| 169 |
+
"""Learnable position encoding used in openai-whisper.decoder"""
|
| 170 |
+
|
| 171 |
+
def __init__(self, d_model: int, dropout_rate: float, max_len: int = 448):
|
| 172 |
+
super().__init__(d_model, dropout_rate, max_len)
|
| 173 |
+
# NOTE(xcsong): overwrite self.pe & self.xscale
|
| 174 |
+
self.pe = torch.nn.Parameter(torch.empty(1, max_len, d_model))
|
| 175 |
+
self.xscale = 1.0
|
| 176 |
+
|
| 177 |
+
|
| 178 |
+
class NoPositionalEncoding(torch.nn.Module):
|
| 179 |
+
"""No position encoding"""
|
| 180 |
+
|
| 181 |
+
def __init__(self, d_model: int, dropout_rate: float):
|
| 182 |
+
super().__init__()
|
| 183 |
+
self.d_model = d_model
|
| 184 |
+
self.dropout = torch.nn.Dropout(p=dropout_rate)
|
| 185 |
+
|
| 186 |
+
def forward(
|
| 187 |
+
self, x: torch.Tensor, offset: Union[int, torch.Tensor] = 0
|
| 188 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 189 |
+
"""Just return zero vector for interface compatibility"""
|
| 190 |
+
pos_emb = torch.zeros(1, x.size(1), self.d_model).to(x.device)
|
| 191 |
+
return self.dropout(x), pos_emb
|
| 192 |
+
|
| 193 |
+
def position_encoding(
|
| 194 |
+
self, offset: Union[int, torch.Tensor], size: int
|
| 195 |
+
) -> torch.Tensor:
|
| 196 |
+
return torch.zeros(1, size, self.d_model)
|
| 197 |
+
|
| 198 |
+
|
| 199 |
+
class EspnetRelPositionalEncoding(torch.nn.Module):
|
| 200 |
+
"""Relative positional encoding module (new implementation).
|
| 201 |
+
|
| 202 |
+
Details can be found in https://github.com/espnet/espnet/pull/2816.
|
| 203 |
+
|
| 204 |
+
See : Appendix B in https://arxiv.org/abs/1901.02860
|
| 205 |
+
|
| 206 |
+
Args:
|
| 207 |
+
d_model (int): Embedding dimension.
|
| 208 |
+
dropout_rate (float): Dropout rate.
|
| 209 |
+
max_len (int): Maximum input length.
|
| 210 |
+
|
| 211 |
+
"""
|
| 212 |
+
|
| 213 |
+
def __init__(self, d_model: int, dropout_rate: float, max_len: int = 5000):
|
| 214 |
+
"""Construct an PositionalEncoding object."""
|
| 215 |
+
super(EspnetRelPositionalEncoding, self).__init__()
|
| 216 |
+
self.d_model = d_model
|
| 217 |
+
self.xscale = math.sqrt(self.d_model)
|
| 218 |
+
self.dropout = torch.nn.Dropout(p=dropout_rate)
|
| 219 |
+
self.pe = None
|
| 220 |
+
self.extend_pe(torch.tensor(0.0).expand(1, max_len))
|
| 221 |
+
|
| 222 |
+
def extend_pe(self, x: torch.Tensor):
|
| 223 |
+
"""Reset the positional encodings."""
|
| 224 |
+
if self.pe is not None:
|
| 225 |
+
# self.pe contains both positive and negative parts
|
| 226 |
+
# the length of self.pe is 2 * input_len - 1
|
| 227 |
+
if self.pe.size(1) >= x.size(1) * 2 - 1:
|
| 228 |
+
if self.pe.dtype != x.dtype or self.pe.device != x.device:
|
| 229 |
+
self.pe = self.pe.to(dtype=x.dtype, device=x.device)
|
| 230 |
+
return
|
| 231 |
+
# Suppose `i` means to the position of query vecotr and `j` means the
|
| 232 |
+
# position of key vector. We use position relative positions when keys
|
| 233 |
+
# are to the left (i>j) and negative relative positions otherwise (i<j).
|
| 234 |
+
pe_positive = torch.zeros(x.size(1), self.d_model)
|
| 235 |
+
pe_negative = torch.zeros(x.size(1), self.d_model)
|
| 236 |
+
position = torch.arange(0, x.size(1), dtype=torch.float32).unsqueeze(1)
|
| 237 |
+
div_term = torch.exp(
|
| 238 |
+
torch.arange(0, self.d_model, 2, dtype=torch.float32)
|
| 239 |
+
* -(math.log(10000.0) / self.d_model)
|
| 240 |
+
)
|
| 241 |
+
pe_positive[:, 0::2] = torch.sin(position * div_term)
|
| 242 |
+
pe_positive[:, 1::2] = torch.cos(position * div_term)
|
| 243 |
+
pe_negative[:, 0::2] = torch.sin(-1 * position * div_term)
|
| 244 |
+
pe_negative[:, 1::2] = torch.cos(-1 * position * div_term)
|
| 245 |
+
|
| 246 |
+
# Reserve the order of positive indices and concat both positive and
|
| 247 |
+
# negative indices. This is used to support the shifting trick
|
| 248 |
+
# as in https://arxiv.org/abs/1901.02860
|
| 249 |
+
pe_positive = torch.flip(pe_positive, [0]).unsqueeze(0)
|
| 250 |
+
pe_negative = pe_negative[1:].unsqueeze(0)
|
| 251 |
+
pe = torch.cat([pe_positive, pe_negative], dim=1)
|
| 252 |
+
self.pe = pe.to(device=x.device, dtype=x.dtype)
|
| 253 |
+
|
| 254 |
+
def forward(
|
| 255 |
+
self, x: torch.Tensor, offset: Union[int, torch.Tensor] = 0
|
| 256 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 257 |
+
"""Add positional encoding.
|
| 258 |
+
|
| 259 |
+
Args:
|
| 260 |
+
x (torch.Tensor): Input tensor (batch, time, `*`).
|
| 261 |
+
|
| 262 |
+
Returns:
|
| 263 |
+
torch.Tensor: Encoded tensor (batch, time, `*`).
|
| 264 |
+
|
| 265 |
+
"""
|
| 266 |
+
self.extend_pe(x)
|
| 267 |
+
x = x * self.xscale
|
| 268 |
+
pos_emb = self.position_encoding(size=x.size(1), offset=offset)
|
| 269 |
+
return self.dropout(x), self.dropout(pos_emb)
|
| 270 |
+
|
| 271 |
+
def position_encoding(
|
| 272 |
+
self, offset: Union[int, torch.Tensor], size: int
|
| 273 |
+
) -> torch.Tensor:
|
| 274 |
+
"""For getting encoding in a streaming fashion
|
| 275 |
+
|
| 276 |
+
Attention!!!!!
|
| 277 |
+
we apply dropout only once at the whole utterance level in a none
|
| 278 |
+
streaming way, but will call this function several times with
|
| 279 |
+
increasing input size in a streaming scenario, so the dropout will
|
| 280 |
+
be applied several times.
|
| 281 |
+
|
| 282 |
+
Args:
|
| 283 |
+
offset (int or torch.tensor): start offset
|
| 284 |
+
size (int): required size of position encoding
|
| 285 |
+
|
| 286 |
+
Returns:
|
| 287 |
+
torch.Tensor: Corresponding encoding
|
| 288 |
+
"""
|
| 289 |
+
pos_emb = self.pe[
|
| 290 |
+
:,
|
| 291 |
+
self.pe.size(1) // 2 - size + 1 : self.pe.size(1) // 2 + size,
|
| 292 |
+
]
|
| 293 |
+
return pos_emb
|
cosyvoice/transformer/encoder.py
ADDED
|
@@ -0,0 +1,633 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2021 Mobvoi Inc (Binbin Zhang, Di Wu)
|
| 2 |
+
# 2022 Xingchen Song ([email protected])
|
| 3 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 4 |
+
#
|
| 5 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 6 |
+
# you may not use this file except in compliance with the License.
|
| 7 |
+
# You may obtain a copy of the License at
|
| 8 |
+
#
|
| 9 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 10 |
+
#
|
| 11 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 12 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 13 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 14 |
+
# See the License for the specific language governing permissions and
|
| 15 |
+
# limitations under the License.
|
| 16 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 17 |
+
"""Encoder definition."""
|
| 18 |
+
from typing import Tuple
|
| 19 |
+
import time
|
| 20 |
+
|
| 21 |
+
import torch
|
| 22 |
+
import torch.utils.checkpoint as ckpt
|
| 23 |
+
import torch.nn.functional as F
|
| 24 |
+
|
| 25 |
+
from cosyvoice.transformer.convolution import ConvolutionModule
|
| 26 |
+
from cosyvoice.transformer.encoder_layer import (
|
| 27 |
+
TransformerEncoderLayer,
|
| 28 |
+
)
|
| 29 |
+
from cosyvoice.transformer.encoder_layer import (
|
| 30 |
+
ConformerEncoderLayer,
|
| 31 |
+
)
|
| 32 |
+
from cosyvoice.transformer.positionwise_feed_forward import (
|
| 33 |
+
PositionwiseFeedForward,
|
| 34 |
+
)
|
| 35 |
+
from cosyvoice.utils.class_utils import (
|
| 36 |
+
COSYVOICE_EMB_CLASSES,
|
| 37 |
+
COSYVOICE_SUBSAMPLE_CLASSES,
|
| 38 |
+
COSYVOICE_ATTENTION_CLASSES,
|
| 39 |
+
COSYVOICE_ACTIVATION_CLASSES,
|
| 40 |
+
)
|
| 41 |
+
from cosyvoice.utils.mask import make_pad_mask
|
| 42 |
+
from cosyvoice.utils.mask import add_optional_chunk_mask
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
class BaseEncoder(torch.nn.Module):
|
| 46 |
+
|
| 47 |
+
def __init__(
|
| 48 |
+
self,
|
| 49 |
+
input_size: int,
|
| 50 |
+
output_size: int = 256,
|
| 51 |
+
attention_heads: int = 4,
|
| 52 |
+
linear_units: int = 2048,
|
| 53 |
+
num_blocks: int = 6,
|
| 54 |
+
dropout_rate: float = 0.1,
|
| 55 |
+
positional_dropout_rate: float = 0.1,
|
| 56 |
+
attention_dropout_rate: float = 0.0,
|
| 57 |
+
input_layer: str = "conv2d",
|
| 58 |
+
pos_enc_layer_type: str = "abs_pos",
|
| 59 |
+
normalize_before: bool = True,
|
| 60 |
+
static_chunk_size: int = 0,
|
| 61 |
+
use_dynamic_chunk: bool = False,
|
| 62 |
+
global_cmvn: torch.nn.Module = None,
|
| 63 |
+
use_dynamic_left_chunk: bool = False,
|
| 64 |
+
gradient_checkpointing: bool = False,
|
| 65 |
+
):
|
| 66 |
+
"""
|
| 67 |
+
Args:
|
| 68 |
+
input_size (int): input dim
|
| 69 |
+
output_size (int): dimension of attention
|
| 70 |
+
attention_heads (int): the number of heads of multi head attention
|
| 71 |
+
linear_units (int): the hidden units number of position-wise feed
|
| 72 |
+
forward
|
| 73 |
+
num_blocks (int): the number of decoder blocks
|
| 74 |
+
dropout_rate (float): dropout rate
|
| 75 |
+
attention_dropout_rate (float): dropout rate in attention
|
| 76 |
+
positional_dropout_rate (float): dropout rate after adding
|
| 77 |
+
positional encoding
|
| 78 |
+
input_layer (str): input layer type.
|
| 79 |
+
optional [linear, conv2d, conv2d6, conv2d8]
|
| 80 |
+
pos_enc_layer_type (str): Encoder positional encoding layer type.
|
| 81 |
+
opitonal [abs_pos, scaled_abs_pos, rel_pos, no_pos]
|
| 82 |
+
normalize_before (bool):
|
| 83 |
+
True: use layer_norm before each sub-block of a layer.
|
| 84 |
+
False: use layer_norm after each sub-block of a layer.
|
| 85 |
+
static_chunk_size (int): chunk size for static chunk training and
|
| 86 |
+
decoding
|
| 87 |
+
use_dynamic_chunk (bool): whether use dynamic chunk size for
|
| 88 |
+
training or not, You can only use fixed chunk(chunk_size > 0)
|
| 89 |
+
or dyanmic chunk size(use_dynamic_chunk = True)
|
| 90 |
+
global_cmvn (Optional[torch.nn.Module]): Optional GlobalCMVN module
|
| 91 |
+
use_dynamic_left_chunk (bool): whether use dynamic left chunk in
|
| 92 |
+
dynamic chunk training
|
| 93 |
+
key_bias: whether use bias in attention.linear_k, False for whisper models.
|
| 94 |
+
gradient_checkpointing: rerunning a forward-pass segment for each
|
| 95 |
+
checkpointed segment during backward.
|
| 96 |
+
"""
|
| 97 |
+
super().__init__()
|
| 98 |
+
self._output_size = output_size
|
| 99 |
+
|
| 100 |
+
self.global_cmvn = global_cmvn
|
| 101 |
+
self.embed = COSYVOICE_SUBSAMPLE_CLASSES[input_layer](
|
| 102 |
+
input_size,
|
| 103 |
+
output_size,
|
| 104 |
+
dropout_rate,
|
| 105 |
+
COSYVOICE_EMB_CLASSES[pos_enc_layer_type](
|
| 106 |
+
output_size, positional_dropout_rate
|
| 107 |
+
),
|
| 108 |
+
)
|
| 109 |
+
|
| 110 |
+
self.normalize_before = normalize_before
|
| 111 |
+
self.after_norm = torch.nn.LayerNorm(output_size, eps=1e-5)
|
| 112 |
+
self.static_chunk_size = static_chunk_size
|
| 113 |
+
self.use_dynamic_chunk = use_dynamic_chunk
|
| 114 |
+
self.use_dynamic_left_chunk = use_dynamic_left_chunk
|
| 115 |
+
self.gradient_checkpointing = gradient_checkpointing
|
| 116 |
+
|
| 117 |
+
def output_size(self) -> int:
|
| 118 |
+
return self._output_size
|
| 119 |
+
|
| 120 |
+
def forward(
|
| 121 |
+
self,
|
| 122 |
+
xs: torch.Tensor,
|
| 123 |
+
xs_lens: torch.Tensor,
|
| 124 |
+
decoding_chunk_size: int = 0,
|
| 125 |
+
num_decoding_left_chunks: int = -1,
|
| 126 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 127 |
+
"""Embed positions in tensor.
|
| 128 |
+
|
| 129 |
+
Args:
|
| 130 |
+
xs: padded input tensor (B, T, D)
|
| 131 |
+
xs_lens: input length (B)
|
| 132 |
+
decoding_chunk_size: decoding chunk size for dynamic chunk
|
| 133 |
+
0: default for training, use random dynamic chunk.
|
| 134 |
+
<0: for decoding, use full chunk.
|
| 135 |
+
>0: for decoding, use fixed chunk size as set.
|
| 136 |
+
num_decoding_left_chunks: number of left chunks, this is for decoding,
|
| 137 |
+
the chunk size is decoding_chunk_size.
|
| 138 |
+
>=0: use num_decoding_left_chunks
|
| 139 |
+
<0: use all left chunks
|
| 140 |
+
Returns:
|
| 141 |
+
encoder output tensor xs, and subsampled masks
|
| 142 |
+
xs: padded output tensor (B, T' ~= T/subsample_rate, D)
|
| 143 |
+
masks: torch.Tensor batch padding mask after subsample
|
| 144 |
+
(B, 1, T' ~= T/subsample_rate)
|
| 145 |
+
NOTE(xcsong):
|
| 146 |
+
We pass the `__call__` method of the modules instead of `forward` to the
|
| 147 |
+
checkpointing API because `__call__` attaches all the hooks of the module.
|
| 148 |
+
https://discuss.pytorch.org/t/any-different-between-model-input-and-model-forward-input/3690/2
|
| 149 |
+
"""
|
| 150 |
+
T = xs.size(1)
|
| 151 |
+
masks = ~make_pad_mask(xs_lens, T).unsqueeze(1) # (B, 1, T)
|
| 152 |
+
if self.global_cmvn is not None:
|
| 153 |
+
xs = self.global_cmvn(xs)
|
| 154 |
+
xs, pos_emb, masks = self.embed(xs, masks)
|
| 155 |
+
mask_pad = masks # (B, 1, T/subsample_rate)
|
| 156 |
+
chunk_masks = add_optional_chunk_mask(
|
| 157 |
+
xs,
|
| 158 |
+
masks,
|
| 159 |
+
self.use_dynamic_chunk,
|
| 160 |
+
self.use_dynamic_left_chunk,
|
| 161 |
+
decoding_chunk_size,
|
| 162 |
+
self.static_chunk_size,
|
| 163 |
+
num_decoding_left_chunks,
|
| 164 |
+
)
|
| 165 |
+
print(f"chunk_masks shape: {chunk_masks.shape}")
|
| 166 |
+
if self.gradient_checkpointing and self.training:
|
| 167 |
+
xs = self.forward_layers_checkpointed(xs, chunk_masks, pos_emb, mask_pad)
|
| 168 |
+
else:
|
| 169 |
+
xs = self.forward_layers(xs, chunk_masks, pos_emb, mask_pad)
|
| 170 |
+
if self.normalize_before:
|
| 171 |
+
xs = self.after_norm(xs)
|
| 172 |
+
# Here we assume the mask is not changed in encoder layers, so just
|
| 173 |
+
# return the masks before encoder layers, and the masks will be used
|
| 174 |
+
# for cross attention with decoder later
|
| 175 |
+
return xs, masks
|
| 176 |
+
|
| 177 |
+
def forward_layers(
|
| 178 |
+
self,
|
| 179 |
+
xs: torch.Tensor,
|
| 180 |
+
chunk_masks: torch.Tensor,
|
| 181 |
+
pos_emb: torch.Tensor,
|
| 182 |
+
mask_pad: torch.Tensor,
|
| 183 |
+
) -> torch.Tensor:
|
| 184 |
+
for layer in self.encoders:
|
| 185 |
+
xs, chunk_masks, _, _ = layer(xs, chunk_masks, pos_emb, mask_pad)
|
| 186 |
+
return xs
|
| 187 |
+
|
| 188 |
+
@torch.jit.unused
|
| 189 |
+
def forward_layers_checkpointed(
|
| 190 |
+
self,
|
| 191 |
+
xs: torch.Tensor,
|
| 192 |
+
chunk_masks: torch.Tensor,
|
| 193 |
+
pos_emb: torch.Tensor,
|
| 194 |
+
mask_pad: torch.Tensor,
|
| 195 |
+
) -> torch.Tensor:
|
| 196 |
+
for layer in self.encoders:
|
| 197 |
+
xs, chunk_masks, _, _ = ckpt.checkpoint(
|
| 198 |
+
layer.__call__, xs, chunk_masks, pos_emb, mask_pad
|
| 199 |
+
)
|
| 200 |
+
return xs
|
| 201 |
+
|
| 202 |
+
@torch.jit.export
|
| 203 |
+
def forward_chunk(
|
| 204 |
+
self,
|
| 205 |
+
xs: torch.Tensor,
|
| 206 |
+
offset: int,
|
| 207 |
+
required_cache_size: int,
|
| 208 |
+
att_cache: torch.Tensor = torch.zeros(0, 0, 0, 0),
|
| 209 |
+
cnn_cache: torch.Tensor = torch.zeros(0, 0, 0, 0),
|
| 210 |
+
att_mask: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 211 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 212 |
+
""" Forward just one chunk
|
| 213 |
+
|
| 214 |
+
Args:
|
| 215 |
+
xs (torch.Tensor): chunk input, with shape (b=1, time, mel-dim),
|
| 216 |
+
where `time == (chunk_size - 1) * subsample_rate + \
|
| 217 |
+
subsample.right_context + 1`
|
| 218 |
+
offset (int): current offset in encoder output time stamp
|
| 219 |
+
required_cache_size (int): cache size required for next chunk
|
| 220 |
+
compuation
|
| 221 |
+
>=0: actual cache size
|
| 222 |
+
<0: means all history cache is required
|
| 223 |
+
att_cache (torch.Tensor): cache tensor for KEY & VALUE in
|
| 224 |
+
transformer/conformer attention, with shape
|
| 225 |
+
(elayers, head, cache_t1, d_k * 2), where
|
| 226 |
+
`head * d_k == hidden-dim` and
|
| 227 |
+
`cache_t1 == chunk_size * num_decoding_left_chunks`.
|
| 228 |
+
cnn_cache (torch.Tensor): cache tensor for cnn_module in conformer,
|
| 229 |
+
(elayers, b=1, hidden-dim, cache_t2), where
|
| 230 |
+
`cache_t2 == cnn.lorder - 1`
|
| 231 |
+
|
| 232 |
+
Returns:
|
| 233 |
+
torch.Tensor: output of current input xs,
|
| 234 |
+
with shape (b=1, chunk_size, hidden-dim).
|
| 235 |
+
torch.Tensor: new attention cache required for next chunk, with
|
| 236 |
+
dynamic shape (elayers, head, ?, d_k * 2)
|
| 237 |
+
depending on required_cache_size.
|
| 238 |
+
torch.Tensor: new conformer cnn cache required for next chunk, with
|
| 239 |
+
same shape as the original cnn_cache.
|
| 240 |
+
|
| 241 |
+
"""
|
| 242 |
+
assert xs.size(0) == 1
|
| 243 |
+
# tmp_masks is just for interface compatibility
|
| 244 |
+
tmp_masks = torch.ones(1, xs.size(1), device=xs.device, dtype=torch.bool)
|
| 245 |
+
tmp_masks = tmp_masks.unsqueeze(1)
|
| 246 |
+
if self.global_cmvn is not None:
|
| 247 |
+
xs = self.global_cmvn(xs)
|
| 248 |
+
# NOTE(xcsong): Before embed, shape(xs) is (b=1, time, mel-dim)
|
| 249 |
+
xs, pos_emb, _ = self.embed(xs, tmp_masks, offset)
|
| 250 |
+
# NOTE(xcsong): After embed, shape(xs) is (b=1, chunk_size, hidden-dim)
|
| 251 |
+
elayers, cache_t1 = att_cache.size(0), att_cache.size(2)
|
| 252 |
+
chunk_size = xs.size(1)
|
| 253 |
+
attention_key_size = cache_t1 + chunk_size
|
| 254 |
+
pos_emb = self.embed.position_encoding(
|
| 255 |
+
offset=offset - cache_t1, size=attention_key_size
|
| 256 |
+
)
|
| 257 |
+
if required_cache_size < 0:
|
| 258 |
+
next_cache_start = 0
|
| 259 |
+
elif required_cache_size == 0:
|
| 260 |
+
next_cache_start = attention_key_size
|
| 261 |
+
else:
|
| 262 |
+
next_cache_start = max(attention_key_size - required_cache_size, 0)
|
| 263 |
+
r_att_cache = []
|
| 264 |
+
r_cnn_cache = []
|
| 265 |
+
for i, layer in enumerate(self.encoders):
|
| 266 |
+
# NOTE(xcsong): Before layer.forward
|
| 267 |
+
# shape(att_cache[i:i + 1]) is (1, head, cache_t1, d_k * 2),
|
| 268 |
+
# shape(cnn_cache[i]) is (b=1, hidden-dim, cache_t2)
|
| 269 |
+
xs, _, new_att_cache, new_cnn_cache = layer(
|
| 270 |
+
xs,
|
| 271 |
+
att_mask,
|
| 272 |
+
pos_emb,
|
| 273 |
+
att_cache=att_cache[i : i + 1] if elayers > 0 else att_cache,
|
| 274 |
+
cnn_cache=cnn_cache[i] if cnn_cache.size(0) > 0 else cnn_cache,
|
| 275 |
+
)
|
| 276 |
+
# NOTE(xcsong): After layer.forward
|
| 277 |
+
# shape(new_att_cache) is (1, head, attention_key_size, d_k * 2),
|
| 278 |
+
# shape(new_cnn_cache) is (b=1, hidden-dim, cache_t2)
|
| 279 |
+
r_att_cache.append(new_att_cache[:, :, next_cache_start:, :])
|
| 280 |
+
r_cnn_cache.append(new_cnn_cache.unsqueeze(0))
|
| 281 |
+
if self.normalize_before:
|
| 282 |
+
xs = self.after_norm(xs)
|
| 283 |
+
|
| 284 |
+
# NOTE(xcsong): shape(r_att_cache) is (elayers, head, ?, d_k * 2),
|
| 285 |
+
# ? may be larger than cache_t1, it depends on required_cache_size
|
| 286 |
+
r_att_cache = torch.cat(r_att_cache, dim=0)
|
| 287 |
+
# NOTE(xcsong): shape(r_cnn_cache) is (e, b=1, hidden-dim, cache_t2)
|
| 288 |
+
r_cnn_cache = torch.cat(r_cnn_cache, dim=0)
|
| 289 |
+
|
| 290 |
+
return (xs, r_att_cache, r_cnn_cache)
|
| 291 |
+
|
| 292 |
+
@torch.jit.unused
|
| 293 |
+
def forward_chunk_by_chunk(
|
| 294 |
+
self,
|
| 295 |
+
xs: torch.Tensor,
|
| 296 |
+
decoding_chunk_size: int,
|
| 297 |
+
num_decoding_left_chunks: int = -1,
|
| 298 |
+
) -> Tuple[torch.Tensor, torch.Tensor]:
|
| 299 |
+
"""Forward input chunk by chunk with chunk_size like a streaming
|
| 300 |
+
fashion
|
| 301 |
+
|
| 302 |
+
Here we should pay special attention to computation cache in the
|
| 303 |
+
streaming style forward chunk by chunk. Three things should be taken
|
| 304 |
+
into account for computation in the current network:
|
| 305 |
+
1. transformer/conformer encoder layers output cache
|
| 306 |
+
2. convolution in conformer
|
| 307 |
+
3. convolution in subsampling
|
| 308 |
+
|
| 309 |
+
However, we don't implement subsampling cache for:
|
| 310 |
+
1. We can control subsampling module to output the right result by
|
| 311 |
+
overlapping input instead of cache left context, even though it
|
| 312 |
+
wastes some computation, but subsampling only takes a very
|
| 313 |
+
small fraction of computation in the whole model.
|
| 314 |
+
2. Typically, there are several covolution layers with subsampling
|
| 315 |
+
in subsampling module, it is tricky and complicated to do cache
|
| 316 |
+
with different convolution layers with different subsampling
|
| 317 |
+
rate.
|
| 318 |
+
3. Currently, nn.Sequential is used to stack all the convolution
|
| 319 |
+
layers in subsampling, we need to rewrite it to make it work
|
| 320 |
+
with cache, which is not preferred.
|
| 321 |
+
Args:
|
| 322 |
+
xs (torch.Tensor): (1, max_len, dim)
|
| 323 |
+
chunk_size (int): decoding chunk size
|
| 324 |
+
"""
|
| 325 |
+
assert decoding_chunk_size > 0
|
| 326 |
+
# The model is trained by static or dynamic chunk
|
| 327 |
+
assert self.static_chunk_size > 0 or self.use_dynamic_chunk
|
| 328 |
+
subsampling = self.embed.subsampling_rate
|
| 329 |
+
context = self.embed.right_context + 1 # Add current frame
|
| 330 |
+
stride = subsampling * decoding_chunk_size
|
| 331 |
+
decoding_window = (decoding_chunk_size - 1) * subsampling + context
|
| 332 |
+
num_frames = xs.size(1)
|
| 333 |
+
att_cache: torch.Tensor = torch.zeros((0, 0, 0, 0), device=xs.device)
|
| 334 |
+
cnn_cache: torch.Tensor = torch.zeros((0, 0, 0, 0), device=xs.device)
|
| 335 |
+
outputs = []
|
| 336 |
+
offset = 0
|
| 337 |
+
required_cache_size = decoding_chunk_size * num_decoding_left_chunks
|
| 338 |
+
|
| 339 |
+
# Feed forward overlap input step by step
|
| 340 |
+
for cur in range(0, num_frames - context + 1, stride):
|
| 341 |
+
end = min(cur + decoding_window, num_frames)
|
| 342 |
+
chunk_xs = xs[:, cur:end, :]
|
| 343 |
+
(y, att_cache, cnn_cache) = self.forward_chunk(
|
| 344 |
+
chunk_xs, offset, required_cache_size, att_cache, cnn_cache
|
| 345 |
+
)
|
| 346 |
+
outputs.append(y)
|
| 347 |
+
offset += y.size(1)
|
| 348 |
+
ys = torch.cat(outputs, 1)
|
| 349 |
+
masks = torch.ones((1, 1, ys.size(1)), device=ys.device, dtype=torch.bool)
|
| 350 |
+
return ys, masks
|
| 351 |
+
|
| 352 |
+
|
| 353 |
+
class TransformerEncoder(BaseEncoder):
|
| 354 |
+
"""Transformer encoder module."""
|
| 355 |
+
|
| 356 |
+
def __init__(
|
| 357 |
+
self,
|
| 358 |
+
input_size: int,
|
| 359 |
+
output_size: int = 256,
|
| 360 |
+
attention_heads: int = 4,
|
| 361 |
+
linear_units: int = 2048,
|
| 362 |
+
num_blocks: int = 6,
|
| 363 |
+
dropout_rate: float = 0.1,
|
| 364 |
+
positional_dropout_rate: float = 0.1,
|
| 365 |
+
attention_dropout_rate: float = 0.0,
|
| 366 |
+
input_layer: str = "conv2d",
|
| 367 |
+
pos_enc_layer_type: str = "abs_pos",
|
| 368 |
+
normalize_before: bool = True,
|
| 369 |
+
static_chunk_size: int = 0,
|
| 370 |
+
use_dynamic_chunk: bool = False,
|
| 371 |
+
global_cmvn: torch.nn.Module = None,
|
| 372 |
+
use_dynamic_left_chunk: bool = False,
|
| 373 |
+
key_bias: bool = True,
|
| 374 |
+
selfattention_layer_type: str = "selfattn",
|
| 375 |
+
activation_type: str = "relu",
|
| 376 |
+
gradient_checkpointing: bool = False,
|
| 377 |
+
):
|
| 378 |
+
"""Construct TransformerEncoder
|
| 379 |
+
|
| 380 |
+
See Encoder for the meaning of each parameter.
|
| 381 |
+
"""
|
| 382 |
+
super().__init__(
|
| 383 |
+
input_size,
|
| 384 |
+
output_size,
|
| 385 |
+
attention_heads,
|
| 386 |
+
linear_units,
|
| 387 |
+
num_blocks,
|
| 388 |
+
dropout_rate,
|
| 389 |
+
positional_dropout_rate,
|
| 390 |
+
attention_dropout_rate,
|
| 391 |
+
input_layer,
|
| 392 |
+
pos_enc_layer_type,
|
| 393 |
+
normalize_before,
|
| 394 |
+
static_chunk_size,
|
| 395 |
+
use_dynamic_chunk,
|
| 396 |
+
global_cmvn,
|
| 397 |
+
use_dynamic_left_chunk,
|
| 398 |
+
gradient_checkpointing,
|
| 399 |
+
)
|
| 400 |
+
activation = COSYVOICE_ACTIVATION_CLASSES[activation_type]()
|
| 401 |
+
self.encoders = torch.nn.ModuleList(
|
| 402 |
+
[
|
| 403 |
+
TransformerEncoderLayer(
|
| 404 |
+
output_size,
|
| 405 |
+
COSYVOICE_ATTENTION_CLASSES[selfattention_layer_type](
|
| 406 |
+
attention_heads, output_size, attention_dropout_rate, key_bias
|
| 407 |
+
),
|
| 408 |
+
PositionwiseFeedForward(
|
| 409 |
+
output_size, linear_units, dropout_rate, activation
|
| 410 |
+
),
|
| 411 |
+
dropout_rate,
|
| 412 |
+
normalize_before,
|
| 413 |
+
)
|
| 414 |
+
for _ in range(num_blocks)
|
| 415 |
+
]
|
| 416 |
+
)
|
| 417 |
+
|
| 418 |
+
|
| 419 |
+
class ConformerEncoder(BaseEncoder):
|
| 420 |
+
"""Conformer encoder module."""
|
| 421 |
+
|
| 422 |
+
def __init__(
|
| 423 |
+
self,
|
| 424 |
+
input_size: int,
|
| 425 |
+
output_size: int = 256,
|
| 426 |
+
attention_heads: int = 4,
|
| 427 |
+
linear_units: int = 2048,
|
| 428 |
+
num_blocks: int = 6,
|
| 429 |
+
dropout_rate: float = 0.1,
|
| 430 |
+
positional_dropout_rate: float = 0.1,
|
| 431 |
+
attention_dropout_rate: float = 0.0,
|
| 432 |
+
input_layer: str = "conv2d",
|
| 433 |
+
pos_enc_layer_type: str = "rel_pos",
|
| 434 |
+
normalize_before: bool = True,
|
| 435 |
+
static_chunk_size: int = 0,
|
| 436 |
+
use_dynamic_chunk: bool = False,
|
| 437 |
+
global_cmvn: torch.nn.Module = None,
|
| 438 |
+
use_dynamic_left_chunk: bool = False,
|
| 439 |
+
positionwise_conv_kernel_size: int = 1,
|
| 440 |
+
macaron_style: bool = True,
|
| 441 |
+
selfattention_layer_type: str = "rel_selfattn",
|
| 442 |
+
activation_type: str = "swish",
|
| 443 |
+
use_cnn_module: bool = True,
|
| 444 |
+
cnn_module_kernel: int = 15,
|
| 445 |
+
causal: bool = False,
|
| 446 |
+
cnn_module_norm: str = "batch_norm",
|
| 447 |
+
key_bias: bool = True,
|
| 448 |
+
gradient_checkpointing: bool = False,
|
| 449 |
+
):
|
| 450 |
+
"""Construct ConformerEncoder
|
| 451 |
+
|
| 452 |
+
Args:
|
| 453 |
+
input_size to use_dynamic_chunk, see in BaseEncoder
|
| 454 |
+
positionwise_conv_kernel_size (int): Kernel size of positionwise
|
| 455 |
+
conv1d layer.
|
| 456 |
+
macaron_style (bool): Whether to use macaron style for
|
| 457 |
+
positionwise layer.
|
| 458 |
+
selfattention_layer_type (str): Encoder attention layer type,
|
| 459 |
+
the parameter has no effect now, it's just for configure
|
| 460 |
+
compatibility.
|
| 461 |
+
activation_type (str): Encoder activation function type.
|
| 462 |
+
use_cnn_module (bool): Whether to use convolution module.
|
| 463 |
+
cnn_module_kernel (int): Kernel size of convolution module.
|
| 464 |
+
causal (bool): whether to use causal convolution or not.
|
| 465 |
+
key_bias: whether use bias in attention.linear_k, False for whisper models.
|
| 466 |
+
"""
|
| 467 |
+
super().__init__(
|
| 468 |
+
input_size,
|
| 469 |
+
output_size,
|
| 470 |
+
attention_heads,
|
| 471 |
+
linear_units,
|
| 472 |
+
num_blocks,
|
| 473 |
+
dropout_rate,
|
| 474 |
+
positional_dropout_rate,
|
| 475 |
+
attention_dropout_rate,
|
| 476 |
+
input_layer,
|
| 477 |
+
pos_enc_layer_type,
|
| 478 |
+
normalize_before,
|
| 479 |
+
static_chunk_size,
|
| 480 |
+
use_dynamic_chunk,
|
| 481 |
+
global_cmvn,
|
| 482 |
+
use_dynamic_left_chunk,
|
| 483 |
+
gradient_checkpointing,
|
| 484 |
+
)
|
| 485 |
+
activation = COSYVOICE_ACTIVATION_CLASSES[activation_type]()
|
| 486 |
+
|
| 487 |
+
# self-attention module definition
|
| 488 |
+
encoder_selfattn_layer_args = (
|
| 489 |
+
attention_heads,
|
| 490 |
+
output_size,
|
| 491 |
+
attention_dropout_rate,
|
| 492 |
+
key_bias,
|
| 493 |
+
)
|
| 494 |
+
# feed-forward module definition
|
| 495 |
+
positionwise_layer_args = (
|
| 496 |
+
output_size,
|
| 497 |
+
linear_units,
|
| 498 |
+
dropout_rate,
|
| 499 |
+
activation,
|
| 500 |
+
)
|
| 501 |
+
# convolution module definition
|
| 502 |
+
convolution_layer_args = (
|
| 503 |
+
output_size,
|
| 504 |
+
cnn_module_kernel,
|
| 505 |
+
activation,
|
| 506 |
+
cnn_module_norm,
|
| 507 |
+
causal,
|
| 508 |
+
)
|
| 509 |
+
|
| 510 |
+
self.encoders = torch.nn.ModuleList(
|
| 511 |
+
[
|
| 512 |
+
ConformerEncoderLayer(
|
| 513 |
+
output_size,
|
| 514 |
+
COSYVOICE_ATTENTION_CLASSES[selfattention_layer_type](
|
| 515 |
+
*encoder_selfattn_layer_args
|
| 516 |
+
),
|
| 517 |
+
PositionwiseFeedForward(*positionwise_layer_args),
|
| 518 |
+
(
|
| 519 |
+
PositionwiseFeedForward(*positionwise_layer_args)
|
| 520 |
+
if macaron_style
|
| 521 |
+
else None
|
| 522 |
+
),
|
| 523 |
+
(
|
| 524 |
+
ConvolutionModule(*convolution_layer_args)
|
| 525 |
+
if use_cnn_module
|
| 526 |
+
else None
|
| 527 |
+
),
|
| 528 |
+
dropout_rate,
|
| 529 |
+
normalize_before,
|
| 530 |
+
)
|
| 531 |
+
for _ in range(num_blocks)
|
| 532 |
+
]
|
| 533 |
+
)
|
| 534 |
+
self.inference_buffers = {}
|
| 535 |
+
self.inference_graphs = {}
|
| 536 |
+
|
| 537 |
+
@torch.inference_mode()
|
| 538 |
+
def capture_inference(self, seq_len_to_capture=[128, 256, 512, 1024]):
|
| 539 |
+
device = next(self.parameters()).device
|
| 540 |
+
start_time = time.time()
|
| 541 |
+
print(
|
| 542 |
+
f"Start capture_inference for ConformerEncoder, seq_len_to_capture: {seq_len_to_capture}"
|
| 543 |
+
)
|
| 544 |
+
|
| 545 |
+
for seq_len in seq_len_to_capture:
|
| 546 |
+
xs = torch.randn(
|
| 547 |
+
1, seq_len, self._output_size, device=device, dtype=torch.bfloat16
|
| 548 |
+
)
|
| 549 |
+
xs_lens = torch.tensor([seq_len], device=device, dtype=torch.int32)
|
| 550 |
+
decoding_chunk_size = 0
|
| 551 |
+
num_decoding_left_chunks = -1
|
| 552 |
+
|
| 553 |
+
T = xs.size(1)
|
| 554 |
+
masks = ~make_pad_mask(xs_lens, T).unsqueeze(1) # (B, 1, T)
|
| 555 |
+
if self.global_cmvn is not None:
|
| 556 |
+
xs = self.global_cmvn(xs)
|
| 557 |
+
xs, pos_emb, masks = self.embed(xs, masks)
|
| 558 |
+
mask_pad = masks # (B, 1, T/subsample_rate)
|
| 559 |
+
chunk_masks = add_optional_chunk_mask(
|
| 560 |
+
xs,
|
| 561 |
+
masks,
|
| 562 |
+
self.use_dynamic_chunk,
|
| 563 |
+
self.use_dynamic_left_chunk,
|
| 564 |
+
decoding_chunk_size,
|
| 565 |
+
self.static_chunk_size,
|
| 566 |
+
num_decoding_left_chunks,
|
| 567 |
+
)
|
| 568 |
+
|
| 569 |
+
g = torch.cuda.CUDAGraph()
|
| 570 |
+
with torch.cuda.graph(g):
|
| 571 |
+
out = self.forward_layers(xs, chunk_masks, pos_emb, mask_pad)
|
| 572 |
+
|
| 573 |
+
self.inference_graphs[seq_len] = g
|
| 574 |
+
self.inference_buffers[seq_len] = {
|
| 575 |
+
"xs": xs,
|
| 576 |
+
"chunk_masks": chunk_masks,
|
| 577 |
+
"pos_emb": pos_emb,
|
| 578 |
+
"mask_pad": mask_pad,
|
| 579 |
+
"out": out,
|
| 580 |
+
}
|
| 581 |
+
end_time = time.time()
|
| 582 |
+
print(
|
| 583 |
+
f"Finish capture_inference for ConformerEncoder, time elapsed: {end_time - start_time}"
|
| 584 |
+
)
|
| 585 |
+
|
| 586 |
+
@torch.inference_mode()
|
| 587 |
+
def inference(self, xs: torch.Tensor, xs_lens: torch.Tensor):
|
| 588 |
+
curr_seq_len = xs.shape[1]
|
| 589 |
+
target_len = None
|
| 590 |
+
|
| 591 |
+
for seq_len in sorted(self.inference_graphs.keys()):
|
| 592 |
+
if seq_len >= curr_seq_len:
|
| 593 |
+
target_len = seq_len
|
| 594 |
+
break
|
| 595 |
+
|
| 596 |
+
if target_len is not None:
|
| 597 |
+
xs = F.pad(xs, (0, 0, 0, target_len - curr_seq_len), "constant", 0)
|
| 598 |
+
|
| 599 |
+
decoding_chunk_size = 0
|
| 600 |
+
num_decoding_left_chunks = -1
|
| 601 |
+
|
| 602 |
+
T = xs.size(1)
|
| 603 |
+
masks = ~make_pad_mask(xs_lens, T).unsqueeze(1) # (B, 1, T)
|
| 604 |
+
if self.global_cmvn is not None:
|
| 605 |
+
xs = self.global_cmvn(xs)
|
| 606 |
+
xs, pos_emb, masks = self.embed(xs, masks)
|
| 607 |
+
mask_pad = masks # (B, 1, T/subsample_rate)
|
| 608 |
+
chunk_masks = add_optional_chunk_mask(
|
| 609 |
+
xs,
|
| 610 |
+
masks,
|
| 611 |
+
self.use_dynamic_chunk,
|
| 612 |
+
self.use_dynamic_left_chunk,
|
| 613 |
+
decoding_chunk_size,
|
| 614 |
+
self.static_chunk_size,
|
| 615 |
+
num_decoding_left_chunks,
|
| 616 |
+
)
|
| 617 |
+
|
| 618 |
+
if target_len is not None:
|
| 619 |
+
buffer = self.inference_buffers[target_len]
|
| 620 |
+
buffer["xs"].copy_(xs)
|
| 621 |
+
buffer["chunk_masks"].copy_(chunk_masks)
|
| 622 |
+
buffer["pos_emb"].copy_(pos_emb)
|
| 623 |
+
buffer["mask_pad"].copy_(mask_pad)
|
| 624 |
+
|
| 625 |
+
self.inference_graphs[target_len].replay()
|
| 626 |
+
|
| 627 |
+
out = buffer["out"][:, :curr_seq_len, :]
|
| 628 |
+
else:
|
| 629 |
+
out = self.forward_layers(xs, chunk_masks, pos_emb, mask_pad)
|
| 630 |
+
|
| 631 |
+
if self.normalize_before:
|
| 632 |
+
out = self.after_norm(out)
|
| 633 |
+
return out, masks
|
cosyvoice/transformer/encoder_layer.py
ADDED
|
@@ -0,0 +1,237 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2021 Mobvoi Inc (Binbin Zhang, Di Wu)
|
| 2 |
+
# 2022 Xingchen Song ([email protected])
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 16 |
+
"""Encoder self-attention layer definition."""
|
| 17 |
+
|
| 18 |
+
from typing import Optional, Tuple
|
| 19 |
+
|
| 20 |
+
import torch
|
| 21 |
+
from torch import nn
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
class TransformerEncoderLayer(nn.Module):
|
| 25 |
+
"""Encoder layer module.
|
| 26 |
+
|
| 27 |
+
Args:
|
| 28 |
+
size (int): Input dimension.
|
| 29 |
+
self_attn (torch.nn.Module): Self-attention module instance.
|
| 30 |
+
`MultiHeadedAttention` or `RelPositionMultiHeadedAttention`
|
| 31 |
+
instance can be used as the argument.
|
| 32 |
+
feed_forward (torch.nn.Module): Feed-forward module instance.
|
| 33 |
+
`PositionwiseFeedForward`, instance can be used as the argument.
|
| 34 |
+
dropout_rate (float): Dropout rate.
|
| 35 |
+
normalize_before (bool):
|
| 36 |
+
True: use layer_norm before each sub-block.
|
| 37 |
+
False: to use layer_norm after each sub-block.
|
| 38 |
+
"""
|
| 39 |
+
|
| 40 |
+
def __init__(
|
| 41 |
+
self,
|
| 42 |
+
size: int,
|
| 43 |
+
self_attn: torch.nn.Module,
|
| 44 |
+
feed_forward: torch.nn.Module,
|
| 45 |
+
dropout_rate: float,
|
| 46 |
+
normalize_before: bool = True,
|
| 47 |
+
):
|
| 48 |
+
"""Construct an EncoderLayer object."""
|
| 49 |
+
super().__init__()
|
| 50 |
+
self.self_attn = self_attn
|
| 51 |
+
self.feed_forward = feed_forward
|
| 52 |
+
self.norm1 = nn.LayerNorm(size, eps=1e-5)
|
| 53 |
+
self.norm2 = nn.LayerNorm(size, eps=1e-5)
|
| 54 |
+
self.dropout = nn.Dropout(dropout_rate)
|
| 55 |
+
self.size = size
|
| 56 |
+
self.normalize_before = normalize_before
|
| 57 |
+
|
| 58 |
+
def forward(
|
| 59 |
+
self,
|
| 60 |
+
x: torch.Tensor,
|
| 61 |
+
mask: torch.Tensor,
|
| 62 |
+
pos_emb: torch.Tensor,
|
| 63 |
+
mask_pad: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 64 |
+
att_cache: torch.Tensor = torch.zeros((0, 0, 0, 0)),
|
| 65 |
+
cnn_cache: torch.Tensor = torch.zeros((0, 0, 0, 0)),
|
| 66 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 67 |
+
"""Compute encoded features.
|
| 68 |
+
|
| 69 |
+
Args:
|
| 70 |
+
x (torch.Tensor): (#batch, time, size)
|
| 71 |
+
mask (torch.Tensor): Mask tensor for the input (#batch, time,time),
|
| 72 |
+
(0, 0, 0) means fake mask.
|
| 73 |
+
pos_emb (torch.Tensor): just for interface compatibility
|
| 74 |
+
to ConformerEncoderLayer
|
| 75 |
+
mask_pad (torch.Tensor): does not used in transformer layer,
|
| 76 |
+
just for unified api with conformer.
|
| 77 |
+
att_cache (torch.Tensor): Cache tensor of the KEY & VALUE
|
| 78 |
+
(#batch=1, head, cache_t1, d_k * 2), head * d_k == size.
|
| 79 |
+
cnn_cache (torch.Tensor): Convolution cache in conformer layer
|
| 80 |
+
(#batch=1, size, cache_t2), not used here, it's for interface
|
| 81 |
+
compatibility to ConformerEncoderLayer.
|
| 82 |
+
Returns:
|
| 83 |
+
torch.Tensor: Output tensor (#batch, time, size).
|
| 84 |
+
torch.Tensor: Mask tensor (#batch, time, time).
|
| 85 |
+
torch.Tensor: att_cache tensor,
|
| 86 |
+
(#batch=1, head, cache_t1 + time, d_k * 2).
|
| 87 |
+
torch.Tensor: cnn_cahce tensor (#batch=1, size, cache_t2).
|
| 88 |
+
|
| 89 |
+
"""
|
| 90 |
+
residual = x
|
| 91 |
+
if self.normalize_before:
|
| 92 |
+
x = self.norm1(x)
|
| 93 |
+
x_att, new_att_cache = self.self_attn(
|
| 94 |
+
x, x, x, mask, pos_emb=pos_emb, cache=att_cache
|
| 95 |
+
)
|
| 96 |
+
x = residual + self.dropout(x_att)
|
| 97 |
+
if not self.normalize_before:
|
| 98 |
+
x = self.norm1(x)
|
| 99 |
+
|
| 100 |
+
residual = x
|
| 101 |
+
if self.normalize_before:
|
| 102 |
+
x = self.norm2(x)
|
| 103 |
+
x = residual + self.dropout(self.feed_forward(x))
|
| 104 |
+
if not self.normalize_before:
|
| 105 |
+
x = self.norm2(x)
|
| 106 |
+
|
| 107 |
+
fake_cnn_cache = torch.zeros((0, 0, 0), dtype=x.dtype, device=x.device)
|
| 108 |
+
return x, mask, new_att_cache, fake_cnn_cache
|
| 109 |
+
|
| 110 |
+
|
| 111 |
+
class ConformerEncoderLayer(nn.Module):
|
| 112 |
+
"""Encoder layer module.
|
| 113 |
+
Args:
|
| 114 |
+
size (int): Input dimension.
|
| 115 |
+
self_attn (torch.nn.Module): Self-attention module instance.
|
| 116 |
+
`MultiHeadedAttention` or `RelPositionMultiHeadedAttention`
|
| 117 |
+
instance can be used as the argument.
|
| 118 |
+
feed_forward (torch.nn.Module): Feed-forward module instance.
|
| 119 |
+
`PositionwiseFeedForward` instance can be used as the argument.
|
| 120 |
+
feed_forward_macaron (torch.nn.Module): Additional feed-forward module
|
| 121 |
+
instance.
|
| 122 |
+
`PositionwiseFeedForward` instance can be used as the argument.
|
| 123 |
+
conv_module (torch.nn.Module): Convolution module instance.
|
| 124 |
+
`ConvlutionModule` instance can be used as the argument.
|
| 125 |
+
dropout_rate (float): Dropout rate.
|
| 126 |
+
normalize_before (bool):
|
| 127 |
+
True: use layer_norm before each sub-block.
|
| 128 |
+
False: use layer_norm after each sub-block.
|
| 129 |
+
"""
|
| 130 |
+
|
| 131 |
+
def __init__(
|
| 132 |
+
self,
|
| 133 |
+
size: int,
|
| 134 |
+
self_attn: torch.nn.Module,
|
| 135 |
+
feed_forward: Optional[nn.Module] = None,
|
| 136 |
+
feed_forward_macaron: Optional[nn.Module] = None,
|
| 137 |
+
conv_module: Optional[nn.Module] = None,
|
| 138 |
+
dropout_rate: float = 0.1,
|
| 139 |
+
normalize_before: bool = True,
|
| 140 |
+
):
|
| 141 |
+
"""Construct an EncoderLayer object."""
|
| 142 |
+
super().__init__()
|
| 143 |
+
self.self_attn = self_attn
|
| 144 |
+
self.feed_forward = feed_forward
|
| 145 |
+
self.feed_forward_macaron = feed_forward_macaron
|
| 146 |
+
self.conv_module = conv_module
|
| 147 |
+
self.norm_ff = nn.LayerNorm(size, eps=1e-5) # for the FNN module
|
| 148 |
+
self.norm_mha = nn.LayerNorm(size, eps=1e-5) # for the MHA module
|
| 149 |
+
if feed_forward_macaron is not None:
|
| 150 |
+
self.norm_ff_macaron = nn.LayerNorm(size, eps=1e-5)
|
| 151 |
+
self.ff_scale = 0.5
|
| 152 |
+
else:
|
| 153 |
+
self.ff_scale = 1.0
|
| 154 |
+
if self.conv_module is not None:
|
| 155 |
+
self.norm_conv = nn.LayerNorm(size, eps=1e-5) # for the CNN module
|
| 156 |
+
self.norm_final = nn.LayerNorm(
|
| 157 |
+
size, eps=1e-5
|
| 158 |
+
) # for the final output of the block
|
| 159 |
+
self.dropout = nn.Dropout(dropout_rate)
|
| 160 |
+
self.size = size
|
| 161 |
+
self.normalize_before = normalize_before
|
| 162 |
+
|
| 163 |
+
def forward(
|
| 164 |
+
self,
|
| 165 |
+
x: torch.Tensor,
|
| 166 |
+
mask: torch.Tensor,
|
| 167 |
+
pos_emb: torch.Tensor,
|
| 168 |
+
mask_pad: torch.Tensor = torch.ones((0, 0, 0), dtype=torch.bool),
|
| 169 |
+
att_cache: torch.Tensor = torch.zeros((0, 0, 0, 0)),
|
| 170 |
+
cnn_cache: torch.Tensor = torch.zeros((0, 0, 0, 0)),
|
| 171 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 172 |
+
"""Compute encoded features.
|
| 173 |
+
|
| 174 |
+
Args:
|
| 175 |
+
x (torch.Tensor): (#batch, time, size)
|
| 176 |
+
mask (torch.Tensor): Mask tensor for the input (#batch, time,time),
|
| 177 |
+
(0, 0, 0) means fake mask.
|
| 178 |
+
pos_emb (torch.Tensor): positional encoding, must not be None
|
| 179 |
+
for ConformerEncoderLayer.
|
| 180 |
+
mask_pad (torch.Tensor): batch padding mask used for conv module.
|
| 181 |
+
(#batch, 1,time), (0, 0, 0) means fake mask.
|
| 182 |
+
att_cache (torch.Tensor): Cache tensor of the KEY & VALUE
|
| 183 |
+
(#batch=1, head, cache_t1, d_k * 2), head * d_k == size.
|
| 184 |
+
cnn_cache (torch.Tensor): Convolution cache in conformer layer
|
| 185 |
+
(#batch=1, size, cache_t2)
|
| 186 |
+
Returns:
|
| 187 |
+
torch.Tensor: Output tensor (#batch, time, size).
|
| 188 |
+
torch.Tensor: Mask tensor (#batch, time, time).
|
| 189 |
+
torch.Tensor: att_cache tensor,
|
| 190 |
+
(#batch=1, head, cache_t1 + time, d_k * 2).
|
| 191 |
+
torch.Tensor: cnn_cahce tensor (#batch, size, cache_t2).
|
| 192 |
+
"""
|
| 193 |
+
|
| 194 |
+
# whether to use macaron style
|
| 195 |
+
if self.feed_forward_macaron is not None:
|
| 196 |
+
residual = x
|
| 197 |
+
if self.normalize_before:
|
| 198 |
+
x = self.norm_ff_macaron(x)
|
| 199 |
+
x = residual + self.ff_scale * self.dropout(self.feed_forward_macaron(x))
|
| 200 |
+
if not self.normalize_before:
|
| 201 |
+
x = self.norm_ff_macaron(x)
|
| 202 |
+
|
| 203 |
+
# multi-headed self-attention module
|
| 204 |
+
residual = x
|
| 205 |
+
if self.normalize_before:
|
| 206 |
+
x = self.norm_mha(x)
|
| 207 |
+
x_att, new_att_cache = self.self_attn(x, x, x, mask, pos_emb, att_cache)
|
| 208 |
+
x = residual + self.dropout(x_att)
|
| 209 |
+
if not self.normalize_before:
|
| 210 |
+
x = self.norm_mha(x)
|
| 211 |
+
|
| 212 |
+
# convolution module
|
| 213 |
+
# Fake new cnn cache here, and then change it in conv_module
|
| 214 |
+
new_cnn_cache = torch.zeros((0, 0, 0), dtype=x.dtype, device=x.device)
|
| 215 |
+
if self.conv_module is not None:
|
| 216 |
+
residual = x
|
| 217 |
+
if self.normalize_before:
|
| 218 |
+
x = self.norm_conv(x)
|
| 219 |
+
x, new_cnn_cache = self.conv_module(x, mask_pad, cnn_cache)
|
| 220 |
+
x = residual + self.dropout(x)
|
| 221 |
+
|
| 222 |
+
if not self.normalize_before:
|
| 223 |
+
x = self.norm_conv(x)
|
| 224 |
+
|
| 225 |
+
# feed forward module
|
| 226 |
+
residual = x
|
| 227 |
+
if self.normalize_before:
|
| 228 |
+
x = self.norm_ff(x)
|
| 229 |
+
|
| 230 |
+
x = residual + self.ff_scale * self.dropout(self.feed_forward(x))
|
| 231 |
+
if not self.normalize_before:
|
| 232 |
+
x = self.norm_ff(x)
|
| 233 |
+
|
| 234 |
+
if self.conv_module is not None:
|
| 235 |
+
x = self.norm_final(x)
|
| 236 |
+
|
| 237 |
+
return x, mask, new_att_cache, new_cnn_cache
|
cosyvoice/transformer/label_smoothing_loss.py
ADDED
|
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2019 Shigeki Karita
|
| 2 |
+
# 2020 Mobvoi Inc (Binbin Zhang)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
"""Label smoothing module."""
|
| 16 |
+
|
| 17 |
+
import torch
|
| 18 |
+
from torch import nn
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
class LabelSmoothingLoss(nn.Module):
|
| 22 |
+
"""Label-smoothing loss.
|
| 23 |
+
|
| 24 |
+
In a standard CE loss, the label's data distribution is:
|
| 25 |
+
[0,1,2] ->
|
| 26 |
+
[
|
| 27 |
+
[1.0, 0.0, 0.0],
|
| 28 |
+
[0.0, 1.0, 0.0],
|
| 29 |
+
[0.0, 0.0, 1.0],
|
| 30 |
+
]
|
| 31 |
+
|
| 32 |
+
In the smoothing version CE Loss,some probabilities
|
| 33 |
+
are taken from the true label prob (1.0) and are divided
|
| 34 |
+
among other labels.
|
| 35 |
+
|
| 36 |
+
e.g.
|
| 37 |
+
smoothing=0.1
|
| 38 |
+
[0,1,2] ->
|
| 39 |
+
[
|
| 40 |
+
[0.9, 0.05, 0.05],
|
| 41 |
+
[0.05, 0.9, 0.05],
|
| 42 |
+
[0.05, 0.05, 0.9],
|
| 43 |
+
]
|
| 44 |
+
|
| 45 |
+
Args:
|
| 46 |
+
size (int): the number of class
|
| 47 |
+
padding_idx (int): padding class id which will be ignored for loss
|
| 48 |
+
smoothing (float): smoothing rate (0.0 means the conventional CE)
|
| 49 |
+
normalize_length (bool):
|
| 50 |
+
normalize loss by sequence length if True
|
| 51 |
+
normalize loss by batch size if False
|
| 52 |
+
"""
|
| 53 |
+
|
| 54 |
+
def __init__(
|
| 55 |
+
self,
|
| 56 |
+
size: int,
|
| 57 |
+
padding_idx: int,
|
| 58 |
+
smoothing: float,
|
| 59 |
+
normalize_length: bool = False,
|
| 60 |
+
):
|
| 61 |
+
"""Construct an LabelSmoothingLoss object."""
|
| 62 |
+
super(LabelSmoothingLoss, self).__init__()
|
| 63 |
+
self.criterion = nn.KLDivLoss(reduction="none")
|
| 64 |
+
self.padding_idx = padding_idx
|
| 65 |
+
self.confidence = 1.0 - smoothing
|
| 66 |
+
self.smoothing = smoothing
|
| 67 |
+
self.size = size
|
| 68 |
+
self.normalize_length = normalize_length
|
| 69 |
+
|
| 70 |
+
def forward(self, x: torch.Tensor, target: torch.Tensor) -> torch.Tensor:
|
| 71 |
+
"""Compute loss between x and target.
|
| 72 |
+
|
| 73 |
+
The model outputs and data labels tensors are flatten to
|
| 74 |
+
(batch*seqlen, class) shape and a mask is applied to the
|
| 75 |
+
padding part which should not be calculated for loss.
|
| 76 |
+
|
| 77 |
+
Args:
|
| 78 |
+
x (torch.Tensor): prediction (batch, seqlen, class)
|
| 79 |
+
target (torch.Tensor):
|
| 80 |
+
target signal masked with self.padding_id (batch, seqlen)
|
| 81 |
+
Returns:
|
| 82 |
+
loss (torch.Tensor) : The KL loss, scalar float value
|
| 83 |
+
"""
|
| 84 |
+
assert x.size(2) == self.size
|
| 85 |
+
batch_size = x.size(0)
|
| 86 |
+
x = x.view(-1, self.size)
|
| 87 |
+
target = target.view(-1)
|
| 88 |
+
# use zeros_like instead of torch.no_grad() for true_dist,
|
| 89 |
+
# since no_grad() can not be exported by JIT
|
| 90 |
+
true_dist = torch.zeros_like(x)
|
| 91 |
+
true_dist.fill_(self.smoothing / (self.size - 1))
|
| 92 |
+
ignore = target == self.padding_idx # (B,)
|
| 93 |
+
total = len(target) - ignore.sum().item()
|
| 94 |
+
target = target.masked_fill(ignore, 0) # avoid -1 index
|
| 95 |
+
true_dist.scatter_(1, target.unsqueeze(1), self.confidence)
|
| 96 |
+
kl = self.criterion(torch.log_softmax(x, dim=1), true_dist)
|
| 97 |
+
denom = total if self.normalize_length else batch_size
|
| 98 |
+
return kl.masked_fill(ignore.unsqueeze(1), 0).sum() / denom
|
cosyvoice/transformer/positionwise_feed_forward.py
ADDED
|
@@ -0,0 +1,116 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2019 Shigeki Karita
|
| 2 |
+
# 2020 Mobvoi Inc (Binbin Zhang)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
"""Positionwise feed forward layer definition."""
|
| 16 |
+
|
| 17 |
+
import torch
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
class PositionwiseFeedForward(torch.nn.Module):
|
| 21 |
+
"""Positionwise feed forward layer.
|
| 22 |
+
|
| 23 |
+
FeedForward are appied on each position of the sequence.
|
| 24 |
+
The output dim is same with the input dim.
|
| 25 |
+
|
| 26 |
+
Args:
|
| 27 |
+
idim (int): Input dimenstion.
|
| 28 |
+
hidden_units (int): The number of hidden units.
|
| 29 |
+
dropout_rate (float): Dropout rate.
|
| 30 |
+
activation (torch.nn.Module): Activation function
|
| 31 |
+
"""
|
| 32 |
+
|
| 33 |
+
def __init__(
|
| 34 |
+
self,
|
| 35 |
+
idim: int,
|
| 36 |
+
hidden_units: int,
|
| 37 |
+
dropout_rate: float,
|
| 38 |
+
activation: torch.nn.Module = torch.nn.ReLU(),
|
| 39 |
+
):
|
| 40 |
+
"""Construct a PositionwiseFeedForward object."""
|
| 41 |
+
super(PositionwiseFeedForward, self).__init__()
|
| 42 |
+
self.w_1 = torch.nn.Linear(idim, hidden_units)
|
| 43 |
+
self.activation = activation
|
| 44 |
+
self.dropout = torch.nn.Dropout(dropout_rate)
|
| 45 |
+
self.w_2 = torch.nn.Linear(hidden_units, idim)
|
| 46 |
+
|
| 47 |
+
def forward(self, xs: torch.Tensor) -> torch.Tensor:
|
| 48 |
+
"""Forward function.
|
| 49 |
+
|
| 50 |
+
Args:
|
| 51 |
+
xs: input tensor (B, L, D)
|
| 52 |
+
Returns:
|
| 53 |
+
output tensor, (B, L, D)
|
| 54 |
+
"""
|
| 55 |
+
return self.w_2(self.dropout(self.activation(self.w_1(xs))))
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
class MoEFFNLayer(torch.nn.Module):
|
| 59 |
+
"""
|
| 60 |
+
Mixture of expert with Positionwise feed forward layer
|
| 61 |
+
See also figure 1 in https://arxiv.org/pdf/2305.15663.pdf
|
| 62 |
+
The output dim is same with the input dim.
|
| 63 |
+
|
| 64 |
+
Modified from https://github.com/Lightning-AI/lit-gpt/pull/823
|
| 65 |
+
https://github.com/mistralai/mistral-src/blob/b46d6/moe_one_file_ref.py#L203-L219
|
| 66 |
+
Args:
|
| 67 |
+
n_expert: number of expert.
|
| 68 |
+
n_expert_per_token: The actual number of experts used for each frame
|
| 69 |
+
idim (int): Input dimenstion.
|
| 70 |
+
hidden_units (int): The number of hidden units.
|
| 71 |
+
dropout_rate (float): Dropout rate.
|
| 72 |
+
activation (torch.nn.Module): Activation function
|
| 73 |
+
"""
|
| 74 |
+
|
| 75 |
+
def __init__(
|
| 76 |
+
self,
|
| 77 |
+
n_expert: int,
|
| 78 |
+
n_expert_per_token: int,
|
| 79 |
+
idim: int,
|
| 80 |
+
hidden_units: int,
|
| 81 |
+
dropout_rate: float,
|
| 82 |
+
activation: torch.nn.Module = torch.nn.ReLU(),
|
| 83 |
+
):
|
| 84 |
+
super(MoEFFNLayer, self).__init__()
|
| 85 |
+
self.gate = torch.nn.Linear(idim, n_expert, bias=False)
|
| 86 |
+
self.experts = torch.nn.ModuleList(
|
| 87 |
+
PositionwiseFeedForward(idim, hidden_units, dropout_rate, activation)
|
| 88 |
+
for _ in range(n_expert)
|
| 89 |
+
)
|
| 90 |
+
self.n_expert_per_token = n_expert_per_token
|
| 91 |
+
|
| 92 |
+
def forward(self, xs: torch.Tensor) -> torch.Tensor:
|
| 93 |
+
"""Foward function.
|
| 94 |
+
Args:
|
| 95 |
+
xs: input tensor (B, L, D)
|
| 96 |
+
Returns:
|
| 97 |
+
output tensor, (B, L, D)
|
| 98 |
+
|
| 99 |
+
"""
|
| 100 |
+
B, L, D = xs.size() # batch size, sequence length, embedding dimension (idim)
|
| 101 |
+
xs = xs.view(-1, D) # (B*L, D)
|
| 102 |
+
router = self.gate(xs) # (B*L, n_expert)
|
| 103 |
+
logits, indices = torch.topk(
|
| 104 |
+
router, self.n_expert_per_token
|
| 105 |
+
) # probs:(B*L, n_expert), indices: (B*L, n_expert)
|
| 106 |
+
weights = torch.nn.functional.softmax(logits, dim=1, dtype=torch.float).to(
|
| 107 |
+
dtype=xs.dtype
|
| 108 |
+
) # (B*L, n_expert_per_token)
|
| 109 |
+
output = torch.zeros_like(xs) # (B*L, D)
|
| 110 |
+
for i, expert in enumerate(self.experts):
|
| 111 |
+
mask = indices == i
|
| 112 |
+
batch_idx, ith_expert = torch.where(mask)
|
| 113 |
+
output[batch_idx] += weights[batch_idx, ith_expert, None] * expert(
|
| 114 |
+
xs[batch_idx]
|
| 115 |
+
)
|
| 116 |
+
return output.view(B, L, D)
|
cosyvoice/transformer/subsampling.py
ADDED
|
@@ -0,0 +1,391 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2021 Mobvoi Inc (Binbin Zhang, Di Wu)
|
| 2 |
+
# 2024 Alibaba Inc (Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 16 |
+
"""Subsampling layer definition."""
|
| 17 |
+
|
| 18 |
+
from typing import Tuple, Union
|
| 19 |
+
|
| 20 |
+
import torch
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
class BaseSubsampling(torch.nn.Module):
|
| 24 |
+
|
| 25 |
+
def __init__(self):
|
| 26 |
+
super().__init__()
|
| 27 |
+
self.right_context = 0
|
| 28 |
+
self.subsampling_rate = 1
|
| 29 |
+
|
| 30 |
+
def position_encoding(
|
| 31 |
+
self, offset: Union[int, torch.Tensor], size: int
|
| 32 |
+
) -> torch.Tensor:
|
| 33 |
+
return self.pos_enc.position_encoding(offset, size)
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
class EmbedinigNoSubsampling(BaseSubsampling):
|
| 37 |
+
"""Embedding input without subsampling"""
|
| 38 |
+
|
| 39 |
+
def __init__(
|
| 40 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 41 |
+
):
|
| 42 |
+
super().__init__()
|
| 43 |
+
self.embed = torch.nn.Embedding(idim, odim)
|
| 44 |
+
self.pos_enc = pos_enc_class
|
| 45 |
+
|
| 46 |
+
def forward(
|
| 47 |
+
self,
|
| 48 |
+
x: torch.Tensor,
|
| 49 |
+
x_mask: torch.Tensor,
|
| 50 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 51 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 52 |
+
"""Input x.
|
| 53 |
+
|
| 54 |
+
Args:
|
| 55 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 56 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 57 |
+
|
| 58 |
+
Returns:
|
| 59 |
+
torch.Tensor: linear input tensor (#batch, time', odim),
|
| 60 |
+
where time' = time .
|
| 61 |
+
torch.Tensor: linear input mask (#batch, 1, time'),
|
| 62 |
+
where time' = time .
|
| 63 |
+
|
| 64 |
+
"""
|
| 65 |
+
x = self.embed(x)
|
| 66 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 67 |
+
return x, pos_emb, x_mask
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
class LinearNoSubsampling(BaseSubsampling):
|
| 71 |
+
"""Linear transform the input without subsampling
|
| 72 |
+
|
| 73 |
+
Args:
|
| 74 |
+
idim (int): Input dimension.
|
| 75 |
+
odim (int): Output dimension.
|
| 76 |
+
dropout_rate (float): Dropout rate.
|
| 77 |
+
|
| 78 |
+
"""
|
| 79 |
+
|
| 80 |
+
def __init__(
|
| 81 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 82 |
+
):
|
| 83 |
+
"""Construct an linear object."""
|
| 84 |
+
super().__init__()
|
| 85 |
+
self.out = torch.nn.Sequential(
|
| 86 |
+
torch.nn.Linear(idim, odim),
|
| 87 |
+
torch.nn.LayerNorm(odim, eps=1e-5),
|
| 88 |
+
torch.nn.Dropout(dropout_rate),
|
| 89 |
+
)
|
| 90 |
+
self.pos_enc = pos_enc_class
|
| 91 |
+
self.right_context = 0
|
| 92 |
+
self.subsampling_rate = 1
|
| 93 |
+
|
| 94 |
+
def forward(
|
| 95 |
+
self,
|
| 96 |
+
x: torch.Tensor,
|
| 97 |
+
x_mask: torch.Tensor,
|
| 98 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 99 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 100 |
+
"""Input x.
|
| 101 |
+
|
| 102 |
+
Args:
|
| 103 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 104 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 105 |
+
|
| 106 |
+
Returns:
|
| 107 |
+
torch.Tensor: linear input tensor (#batch, time', odim),
|
| 108 |
+
where time' = time .
|
| 109 |
+
torch.Tensor: linear input mask (#batch, 1, time'),
|
| 110 |
+
where time' = time .
|
| 111 |
+
|
| 112 |
+
"""
|
| 113 |
+
x = self.out(x)
|
| 114 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 115 |
+
return x, pos_emb, x_mask
|
| 116 |
+
|
| 117 |
+
|
| 118 |
+
class Conv1dSubsampling2(BaseSubsampling):
|
| 119 |
+
"""Convolutional 1D subsampling (to 1/2 length).
|
| 120 |
+
It is designed for Whisper, ref:
|
| 121 |
+
https://github.com/openai/whisper/blob/main/whisper/model.py
|
| 122 |
+
|
| 123 |
+
Args:
|
| 124 |
+
idim (int): Input dimension.
|
| 125 |
+
odim (int): Output dimension.
|
| 126 |
+
dropout_rate (float): Dropout rate.
|
| 127 |
+
|
| 128 |
+
"""
|
| 129 |
+
|
| 130 |
+
def __init__(
|
| 131 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 132 |
+
):
|
| 133 |
+
"""Construct an Conv1dSubsampling2 object."""
|
| 134 |
+
super().__init__()
|
| 135 |
+
self.conv = torch.nn.Sequential(
|
| 136 |
+
torch.nn.Conv1d(idim, odim, kernel_size=3, padding=1),
|
| 137 |
+
torch.nn.GELU(),
|
| 138 |
+
torch.nn.Conv1d(odim, odim, kernel_size=3, stride=2, padding=1),
|
| 139 |
+
torch.nn.GELU(),
|
| 140 |
+
)
|
| 141 |
+
self.pos_enc = pos_enc_class
|
| 142 |
+
# The right context for every conv layer is computed by:
|
| 143 |
+
# (kernel_size - 1) * frame_rate_of_this_layer
|
| 144 |
+
self.subsampling_rate = 2
|
| 145 |
+
# 4 = (3 - 1) * 1 + (3 - 1) * 1
|
| 146 |
+
self.right_context = 4
|
| 147 |
+
|
| 148 |
+
def forward(
|
| 149 |
+
self,
|
| 150 |
+
x: torch.Tensor,
|
| 151 |
+
x_mask: torch.Tensor,
|
| 152 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 153 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 154 |
+
"""Subsample x.
|
| 155 |
+
|
| 156 |
+
Args:
|
| 157 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 158 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 159 |
+
|
| 160 |
+
Returns:
|
| 161 |
+
torch.Tensor: Subsampled tensor (#batch, time', odim),
|
| 162 |
+
where time' = time // 2.
|
| 163 |
+
torch.Tensor: Subsampled mask (#batch, 1, time'),
|
| 164 |
+
where time' = time // 2.
|
| 165 |
+
torch.Tensor: positional encoding
|
| 166 |
+
|
| 167 |
+
"""
|
| 168 |
+
time = x.size(1)
|
| 169 |
+
x = x.transpose(1, 2) # (b, f, t)
|
| 170 |
+
x = self.conv(x)
|
| 171 |
+
x = x.transpose(1, 2) # (b, t, f)
|
| 172 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 173 |
+
return x, pos_emb, x_mask[:, :, (time + 1) % 2 :: 2]
|
| 174 |
+
|
| 175 |
+
|
| 176 |
+
class Conv2dSubsampling4(BaseSubsampling):
|
| 177 |
+
"""Convolutional 2D subsampling (to 1/4 length).
|
| 178 |
+
|
| 179 |
+
Args:
|
| 180 |
+
idim (int): Input dimension.
|
| 181 |
+
odim (int): Output dimension.
|
| 182 |
+
dropout_rate (float): Dropout rate.
|
| 183 |
+
|
| 184 |
+
"""
|
| 185 |
+
|
| 186 |
+
def __init__(
|
| 187 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 188 |
+
):
|
| 189 |
+
"""Construct an Conv2dSubsampling4 object."""
|
| 190 |
+
super().__init__()
|
| 191 |
+
self.conv = torch.nn.Sequential(
|
| 192 |
+
torch.nn.Conv2d(1, odim, 3, 2),
|
| 193 |
+
torch.nn.ReLU(),
|
| 194 |
+
torch.nn.Conv2d(odim, odim, 3, 2),
|
| 195 |
+
torch.nn.ReLU(),
|
| 196 |
+
)
|
| 197 |
+
self.out = torch.nn.Sequential(
|
| 198 |
+
torch.nn.Linear(odim * (((idim - 1) // 2 - 1) // 2), odim)
|
| 199 |
+
)
|
| 200 |
+
self.pos_enc = pos_enc_class
|
| 201 |
+
# The right context for every conv layer is computed by:
|
| 202 |
+
# (kernel_size - 1) * frame_rate_of_this_layer
|
| 203 |
+
self.subsampling_rate = 4
|
| 204 |
+
# 6 = (3 - 1) * 1 + (3 - 1) * 2
|
| 205 |
+
self.right_context = 6
|
| 206 |
+
|
| 207 |
+
def forward(
|
| 208 |
+
self,
|
| 209 |
+
x: torch.Tensor,
|
| 210 |
+
x_mask: torch.Tensor,
|
| 211 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 212 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 213 |
+
"""Subsample x.
|
| 214 |
+
|
| 215 |
+
Args:
|
| 216 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 217 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 218 |
+
|
| 219 |
+
Returns:
|
| 220 |
+
torch.Tensor: Subsampled tensor (#batch, time', odim),
|
| 221 |
+
where time' = time // 4.
|
| 222 |
+
torch.Tensor: Subsampled mask (#batch, 1, time'),
|
| 223 |
+
where time' = time // 4.
|
| 224 |
+
torch.Tensor: positional encoding
|
| 225 |
+
|
| 226 |
+
"""
|
| 227 |
+
x = x.unsqueeze(1) # (b, c=1, t, f)
|
| 228 |
+
x = self.conv(x)
|
| 229 |
+
b, c, t, f = x.size()
|
| 230 |
+
x = self.out(x.transpose(1, 2).contiguous().view(b, t, c * f))
|
| 231 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 232 |
+
return x, pos_emb, x_mask[:, :, 2::2][:, :, 2::2]
|
| 233 |
+
|
| 234 |
+
|
| 235 |
+
class Conv2dSubsampling6(BaseSubsampling):
|
| 236 |
+
"""Convolutional 2D subsampling (to 1/6 length).
|
| 237 |
+
Args:
|
| 238 |
+
idim (int): Input dimension.
|
| 239 |
+
odim (int): Output dimension.
|
| 240 |
+
dropout_rate (float): Dropout rate.
|
| 241 |
+
pos_enc (torch.nn.Module): Custom position encoding layer.
|
| 242 |
+
"""
|
| 243 |
+
|
| 244 |
+
def __init__(
|
| 245 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 246 |
+
):
|
| 247 |
+
"""Construct an Conv2dSubsampling6 object."""
|
| 248 |
+
super().__init__()
|
| 249 |
+
self.conv = torch.nn.Sequential(
|
| 250 |
+
torch.nn.Conv2d(1, odim, 3, 2),
|
| 251 |
+
torch.nn.ReLU(),
|
| 252 |
+
torch.nn.Conv2d(odim, odim, 5, 3),
|
| 253 |
+
torch.nn.ReLU(),
|
| 254 |
+
)
|
| 255 |
+
self.linear = torch.nn.Linear(odim * (((idim - 1) // 2 - 2) // 3), odim)
|
| 256 |
+
self.pos_enc = pos_enc_class
|
| 257 |
+
# 10 = (3 - 1) * 1 + (5 - 1) * 2
|
| 258 |
+
self.subsampling_rate = 6
|
| 259 |
+
self.right_context = 10
|
| 260 |
+
|
| 261 |
+
def forward(
|
| 262 |
+
self,
|
| 263 |
+
x: torch.Tensor,
|
| 264 |
+
x_mask: torch.Tensor,
|
| 265 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 266 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 267 |
+
"""Subsample x.
|
| 268 |
+
Args:
|
| 269 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 270 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 271 |
+
|
| 272 |
+
Returns:
|
| 273 |
+
torch.Tensor: Subsampled tensor (#batch, time', odim),
|
| 274 |
+
where time' = time // 6.
|
| 275 |
+
torch.Tensor: Subsampled mask (#batch, 1, time'),
|
| 276 |
+
where time' = time // 6.
|
| 277 |
+
torch.Tensor: positional encoding
|
| 278 |
+
"""
|
| 279 |
+
x = x.unsqueeze(1) # (b, c, t, f)
|
| 280 |
+
x = self.conv(x)
|
| 281 |
+
b, c, t, f = x.size()
|
| 282 |
+
x = self.linear(x.transpose(1, 2).contiguous().view(b, t, c * f))
|
| 283 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 284 |
+
return x, pos_emb, x_mask[:, :, 2::2][:, :, 4::3]
|
| 285 |
+
|
| 286 |
+
|
| 287 |
+
class Conv2dSubsampling8(BaseSubsampling):
|
| 288 |
+
"""Convolutional 2D subsampling (to 1/8 length).
|
| 289 |
+
|
| 290 |
+
Args:
|
| 291 |
+
idim (int): Input dimension.
|
| 292 |
+
odim (int): Output dimension.
|
| 293 |
+
dropout_rate (float): Dropout rate.
|
| 294 |
+
|
| 295 |
+
"""
|
| 296 |
+
|
| 297 |
+
def __init__(
|
| 298 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 299 |
+
):
|
| 300 |
+
"""Construct an Conv2dSubsampling8 object."""
|
| 301 |
+
super().__init__()
|
| 302 |
+
self.conv = torch.nn.Sequential(
|
| 303 |
+
torch.nn.Conv2d(1, odim, 3, 2),
|
| 304 |
+
torch.nn.ReLU(),
|
| 305 |
+
torch.nn.Conv2d(odim, odim, 3, 2),
|
| 306 |
+
torch.nn.ReLU(),
|
| 307 |
+
torch.nn.Conv2d(odim, odim, 3, 2),
|
| 308 |
+
torch.nn.ReLU(),
|
| 309 |
+
)
|
| 310 |
+
self.linear = torch.nn.Linear(
|
| 311 |
+
odim * ((((idim - 1) // 2 - 1) // 2 - 1) // 2), odim
|
| 312 |
+
)
|
| 313 |
+
self.pos_enc = pos_enc_class
|
| 314 |
+
self.subsampling_rate = 8
|
| 315 |
+
# 14 = (3 - 1) * 1 + (3 - 1) * 2 + (3 - 1) * 4
|
| 316 |
+
self.right_context = 14
|
| 317 |
+
|
| 318 |
+
def forward(
|
| 319 |
+
self,
|
| 320 |
+
x: torch.Tensor,
|
| 321 |
+
x_mask: torch.Tensor,
|
| 322 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 323 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 324 |
+
"""Subsample x.
|
| 325 |
+
|
| 326 |
+
Args:
|
| 327 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 328 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 329 |
+
|
| 330 |
+
Returns:
|
| 331 |
+
torch.Tensor: Subsampled tensor (#batch, time', odim),
|
| 332 |
+
where time' = time // 8.
|
| 333 |
+
torch.Tensor: Subsampled mask (#batch, 1, time'),
|
| 334 |
+
where time' = time // 8.
|
| 335 |
+
torch.Tensor: positional encoding
|
| 336 |
+
"""
|
| 337 |
+
x = x.unsqueeze(1) # (b, c, t, f)
|
| 338 |
+
x = self.conv(x)
|
| 339 |
+
b, c, t, f = x.size()
|
| 340 |
+
x = self.linear(x.transpose(1, 2).contiguous().view(b, t, c * f))
|
| 341 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 342 |
+
return x, pos_emb, x_mask[:, :, 2::2][:, :, 2::2][:, :, 2::2]
|
| 343 |
+
|
| 344 |
+
|
| 345 |
+
class LegacyLinearNoSubsampling(BaseSubsampling):
|
| 346 |
+
"""Linear transform the input without subsampling
|
| 347 |
+
|
| 348 |
+
Args:
|
| 349 |
+
idim (int): Input dimension.
|
| 350 |
+
odim (int): Output dimension.
|
| 351 |
+
dropout_rate (float): Dropout rate.
|
| 352 |
+
|
| 353 |
+
"""
|
| 354 |
+
|
| 355 |
+
def __init__(
|
| 356 |
+
self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module
|
| 357 |
+
):
|
| 358 |
+
"""Construct an linear object."""
|
| 359 |
+
super().__init__()
|
| 360 |
+
self.out = torch.nn.Sequential(
|
| 361 |
+
torch.nn.Linear(idim, odim),
|
| 362 |
+
torch.nn.LayerNorm(odim, eps=1e-5),
|
| 363 |
+
torch.nn.Dropout(dropout_rate),
|
| 364 |
+
torch.nn.ReLU(),
|
| 365 |
+
)
|
| 366 |
+
self.pos_enc = pos_enc_class
|
| 367 |
+
self.right_context = 0
|
| 368 |
+
self.subsampling_rate = 1
|
| 369 |
+
|
| 370 |
+
def forward(
|
| 371 |
+
self,
|
| 372 |
+
x: torch.Tensor,
|
| 373 |
+
x_mask: torch.Tensor,
|
| 374 |
+
offset: Union[int, torch.Tensor] = 0,
|
| 375 |
+
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
|
| 376 |
+
"""Input x.
|
| 377 |
+
|
| 378 |
+
Args:
|
| 379 |
+
x (torch.Tensor): Input tensor (#batch, time, idim).
|
| 380 |
+
x_mask (torch.Tensor): Input mask (#batch, 1, time).
|
| 381 |
+
|
| 382 |
+
Returns:
|
| 383 |
+
torch.Tensor: linear input tensor (#batch, time', odim),
|
| 384 |
+
where time' = time .
|
| 385 |
+
torch.Tensor: linear input mask (#batch, 1, time'),
|
| 386 |
+
where time' = time .
|
| 387 |
+
|
| 388 |
+
"""
|
| 389 |
+
x = self.out(x)
|
| 390 |
+
x, pos_emb = self.pos_enc(x, offset)
|
| 391 |
+
return x, pos_emb, x_mask
|
cosyvoice/utils/__init__.py
ADDED
|
File without changes
|
cosyvoice/utils/audio.py
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import numpy as np
|
| 2 |
+
import torch
|
| 3 |
+
import torch.utils.data
|
| 4 |
+
from librosa.filters import mel as librosa_mel_fn
|
| 5 |
+
from scipy.io.wavfile import read
|
| 6 |
+
|
| 7 |
+
MAX_WAV_VALUE = 32768.0
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def load_wav(full_path):
|
| 11 |
+
sampling_rate, data = read(full_path)
|
| 12 |
+
return data, sampling_rate
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def dynamic_range_compression(x, C=1, clip_val=1e-5):
|
| 16 |
+
return np.log(np.clip(x, a_min=clip_val, a_max=None) * C)
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
def dynamic_range_decompression(x, C=1):
|
| 20 |
+
return np.exp(x) / C
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
def dynamic_range_compression_torch(x, C=1, clip_val=1e-5):
|
| 24 |
+
return torch.log(torch.clamp(x, min=clip_val) * C)
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
def dynamic_range_decompression_torch(x, C=1):
|
| 28 |
+
return torch.exp(x) / C
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
def spectral_normalize_torch(magnitudes):
|
| 32 |
+
output = dynamic_range_compression_torch(magnitudes)
|
| 33 |
+
return output
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
def spectral_de_normalize_torch(magnitudes):
|
| 37 |
+
output = dynamic_range_decompression_torch(magnitudes)
|
| 38 |
+
return output
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
mel_basis = {}
|
| 42 |
+
hann_window = {}
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
def mel_spectrogram(
|
| 46 |
+
y, n_fft, num_mels, sampling_rate, hop_size, win_size, fmin, fmax, center=False
|
| 47 |
+
):
|
| 48 |
+
# if torch.min(y) < -1.0:
|
| 49 |
+
# print("min value is ", torch.min(y))
|
| 50 |
+
# if torch.max(y) > 1.0:
|
| 51 |
+
# print("max value is ", torch.max(y))
|
| 52 |
+
|
| 53 |
+
global mel_basis, hann_window # pylint: disable=global-statement
|
| 54 |
+
if f"{str(fmax)}_{str(y.device)}" not in mel_basis:
|
| 55 |
+
mel = librosa_mel_fn(
|
| 56 |
+
sr=sampling_rate, n_fft=n_fft, n_mels=num_mels, fmin=fmin, fmax=fmax
|
| 57 |
+
)
|
| 58 |
+
mel_basis[str(fmax) + "_" + str(y.device)] = (
|
| 59 |
+
torch.from_numpy(mel).float().to(y.device)
|
| 60 |
+
)
|
| 61 |
+
hann_window[str(y.device)] = torch.hann_window(win_size).to(y.device)
|
| 62 |
+
|
| 63 |
+
y = torch.nn.functional.pad(
|
| 64 |
+
y.unsqueeze(1),
|
| 65 |
+
(int((n_fft - hop_size) / 2), int((n_fft - hop_size) / 2)),
|
| 66 |
+
mode="reflect",
|
| 67 |
+
)
|
| 68 |
+
y = y.squeeze(1)
|
| 69 |
+
|
| 70 |
+
spec = torch.view_as_real(
|
| 71 |
+
torch.stft(
|
| 72 |
+
y,
|
| 73 |
+
n_fft,
|
| 74 |
+
hop_length=hop_size,
|
| 75 |
+
win_length=win_size,
|
| 76 |
+
window=hann_window[str(y.device)],
|
| 77 |
+
center=center,
|
| 78 |
+
pad_mode="reflect",
|
| 79 |
+
normalized=False,
|
| 80 |
+
onesided=True,
|
| 81 |
+
return_complex=True,
|
| 82 |
+
)
|
| 83 |
+
)
|
| 84 |
+
|
| 85 |
+
spec = torch.sqrt(spec.pow(2).sum(-1) + (1e-9))
|
| 86 |
+
|
| 87 |
+
spec = torch.matmul(mel_basis[str(fmax) + "_" + str(y.device)], spec)
|
| 88 |
+
spec = spectral_normalize_torch(spec)
|
| 89 |
+
|
| 90 |
+
return spec
|
cosyvoice/utils/class_utils.py
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright [2023-11-28] <[email protected], Xingchen Song>
|
| 2 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
import torch
|
| 16 |
+
|
| 17 |
+
from cosyvoice.transformer.activation import Swish
|
| 18 |
+
from cosyvoice.transformer.subsampling import (
|
| 19 |
+
LinearNoSubsampling,
|
| 20 |
+
EmbedinigNoSubsampling,
|
| 21 |
+
Conv1dSubsampling2,
|
| 22 |
+
Conv2dSubsampling4,
|
| 23 |
+
Conv2dSubsampling6,
|
| 24 |
+
Conv2dSubsampling8,
|
| 25 |
+
)
|
| 26 |
+
from cosyvoice.transformer.embedding import (
|
| 27 |
+
PositionalEncoding,
|
| 28 |
+
RelPositionalEncoding,
|
| 29 |
+
WhisperPositionalEncoding,
|
| 30 |
+
LearnablePositionalEncoding,
|
| 31 |
+
NoPositionalEncoding,
|
| 32 |
+
)
|
| 33 |
+
from cosyvoice.transformer.attention import (
|
| 34 |
+
MultiHeadedAttention,
|
| 35 |
+
RelPositionMultiHeadedAttention,
|
| 36 |
+
)
|
| 37 |
+
from cosyvoice.transformer.embedding import (
|
| 38 |
+
EspnetRelPositionalEncoding,
|
| 39 |
+
)
|
| 40 |
+
from cosyvoice.transformer.subsampling import (
|
| 41 |
+
LegacyLinearNoSubsampling,
|
| 42 |
+
)
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
COSYVOICE_ACTIVATION_CLASSES = {
|
| 46 |
+
"hardtanh": torch.nn.Hardtanh,
|
| 47 |
+
"tanh": torch.nn.Tanh,
|
| 48 |
+
"relu": torch.nn.ReLU,
|
| 49 |
+
"selu": torch.nn.SELU,
|
| 50 |
+
"swish": getattr(torch.nn, "SiLU", Swish),
|
| 51 |
+
"gelu": torch.nn.GELU,
|
| 52 |
+
}
|
| 53 |
+
|
| 54 |
+
COSYVOICE_SUBSAMPLE_CLASSES = {
|
| 55 |
+
"linear": LinearNoSubsampling,
|
| 56 |
+
"linear_legacy": LegacyLinearNoSubsampling,
|
| 57 |
+
"embed": EmbedinigNoSubsampling,
|
| 58 |
+
"conv1d2": Conv1dSubsampling2,
|
| 59 |
+
"conv2d": Conv2dSubsampling4,
|
| 60 |
+
"conv2d6": Conv2dSubsampling6,
|
| 61 |
+
"conv2d8": Conv2dSubsampling8,
|
| 62 |
+
"paraformer_dummy": torch.nn.Identity,
|
| 63 |
+
}
|
| 64 |
+
|
| 65 |
+
COSYVOICE_EMB_CLASSES = {
|
| 66 |
+
"embed": PositionalEncoding,
|
| 67 |
+
"abs_pos": PositionalEncoding,
|
| 68 |
+
"rel_pos": RelPositionalEncoding,
|
| 69 |
+
"rel_pos_espnet": EspnetRelPositionalEncoding,
|
| 70 |
+
"no_pos": NoPositionalEncoding,
|
| 71 |
+
"abs_pos_whisper": WhisperPositionalEncoding,
|
| 72 |
+
"embed_learnable_pe": LearnablePositionalEncoding,
|
| 73 |
+
}
|
| 74 |
+
|
| 75 |
+
COSYVOICE_ATTENTION_CLASSES = {
|
| 76 |
+
"selfattn": MultiHeadedAttention,
|
| 77 |
+
"rel_selfattn": RelPositionMultiHeadedAttention,
|
| 78 |
+
}
|
cosyvoice/utils/common.py
ADDED
|
@@ -0,0 +1,169 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2020 Mobvoi Inc (Binbin Zhang)
|
| 2 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 16 |
+
"""Unility functions for Transformer."""
|
| 17 |
+
|
| 18 |
+
import random
|
| 19 |
+
from typing import List
|
| 20 |
+
|
| 21 |
+
import numpy as np
|
| 22 |
+
import torch
|
| 23 |
+
|
| 24 |
+
IGNORE_ID = -1
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
def pad_list(xs: List[torch.Tensor], pad_value: int):
|
| 28 |
+
"""Perform padding for the list of tensors.
|
| 29 |
+
|
| 30 |
+
Args:
|
| 31 |
+
xs (List): List of Tensors [(T_1, `*`), (T_2, `*`), ..., (T_B, `*`)].
|
| 32 |
+
pad_value (float): Value for padding.
|
| 33 |
+
|
| 34 |
+
Returns:
|
| 35 |
+
Tensor: Padded tensor (B, Tmax, `*`).
|
| 36 |
+
|
| 37 |
+
Examples:
|
| 38 |
+
>>> x = [torch.ones(4), torch.ones(2), torch.ones(1)]
|
| 39 |
+
>>> x
|
| 40 |
+
[tensor([1., 1., 1., 1.]), tensor([1., 1.]), tensor([1.])]
|
| 41 |
+
>>> pad_list(x, 0)
|
| 42 |
+
tensor([[1., 1., 1., 1.],
|
| 43 |
+
[1., 1., 0., 0.],
|
| 44 |
+
[1., 0., 0., 0.]])
|
| 45 |
+
|
| 46 |
+
"""
|
| 47 |
+
max_len = max([len(item) for item in xs])
|
| 48 |
+
batchs = len(xs)
|
| 49 |
+
ndim = xs[0].ndim
|
| 50 |
+
if ndim == 1:
|
| 51 |
+
pad_res = torch.zeros(batchs, max_len, dtype=xs[0].dtype, device=xs[0].device)
|
| 52 |
+
elif ndim == 2:
|
| 53 |
+
pad_res = torch.zeros(
|
| 54 |
+
batchs, max_len, xs[0].shape[1], dtype=xs[0].dtype, device=xs[0].device
|
| 55 |
+
)
|
| 56 |
+
elif ndim == 3:
|
| 57 |
+
pad_res = torch.zeros(
|
| 58 |
+
batchs,
|
| 59 |
+
max_len,
|
| 60 |
+
xs[0].shape[1],
|
| 61 |
+
xs[0].shape[2],
|
| 62 |
+
dtype=xs[0].dtype,
|
| 63 |
+
device=xs[0].device,
|
| 64 |
+
)
|
| 65 |
+
else:
|
| 66 |
+
raise ValueError(f"Unsupported ndim: {ndim}")
|
| 67 |
+
pad_res.fill_(pad_value)
|
| 68 |
+
for i in range(batchs):
|
| 69 |
+
pad_res[i, : len(xs[i])] = xs[i]
|
| 70 |
+
return pad_res
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
def th_accuracy(
|
| 74 |
+
pad_outputs: torch.Tensor, pad_targets: torch.Tensor, ignore_label: int
|
| 75 |
+
) -> torch.Tensor:
|
| 76 |
+
"""Calculate accuracy.
|
| 77 |
+
|
| 78 |
+
Args:
|
| 79 |
+
pad_outputs (Tensor): Prediction tensors (B * Lmax, D).
|
| 80 |
+
pad_targets (LongTensor): Target label tensors (B, Lmax).
|
| 81 |
+
ignore_label (int): Ignore label id.
|
| 82 |
+
|
| 83 |
+
Returns:
|
| 84 |
+
torch.Tensor: Accuracy value (0.0 - 1.0).
|
| 85 |
+
|
| 86 |
+
"""
|
| 87 |
+
pad_pred = pad_outputs.view(
|
| 88 |
+
pad_targets.size(0), pad_targets.size(1), pad_outputs.size(1)
|
| 89 |
+
).argmax(2)
|
| 90 |
+
mask = pad_targets != ignore_label
|
| 91 |
+
numerator = torch.sum(
|
| 92 |
+
pad_pred.masked_select(mask) == pad_targets.masked_select(mask)
|
| 93 |
+
)
|
| 94 |
+
denominator = torch.sum(mask)
|
| 95 |
+
return (numerator / denominator).detach()
|
| 96 |
+
|
| 97 |
+
|
| 98 |
+
def get_padding(kernel_size, dilation=1):
|
| 99 |
+
return int((kernel_size * dilation - dilation) / 2)
|
| 100 |
+
|
| 101 |
+
|
| 102 |
+
def init_weights(m, mean=0.0, std=0.01):
|
| 103 |
+
classname = m.__class__.__name__
|
| 104 |
+
if classname.find("Conv") != -1:
|
| 105 |
+
m.weight.data.normal_(mean, std)
|
| 106 |
+
|
| 107 |
+
|
| 108 |
+
# Repetition Aware Sampling in VALL-E 2
|
| 109 |
+
def ras_sampling(
|
| 110 |
+
weighted_scores,
|
| 111 |
+
decoded_tokens,
|
| 112 |
+
sampling,
|
| 113 |
+
top_p=0.8,
|
| 114 |
+
top_k=25,
|
| 115 |
+
win_size=10,
|
| 116 |
+
tau_r=0.1,
|
| 117 |
+
):
|
| 118 |
+
top_ids = nucleus_sampling(weighted_scores, top_p=top_p, top_k=top_k)
|
| 119 |
+
rep_num = (
|
| 120 |
+
(torch.tensor(decoded_tokens[-win_size:]).to(weighted_scores.device) == top_ids)
|
| 121 |
+
.sum()
|
| 122 |
+
.item()
|
| 123 |
+
)
|
| 124 |
+
if rep_num >= win_size * tau_r:
|
| 125 |
+
top_ids = random_sampling(weighted_scores, decoded_tokens, sampling)
|
| 126 |
+
return top_ids
|
| 127 |
+
|
| 128 |
+
|
| 129 |
+
def nucleus_sampling(weighted_scores, top_p=0.8, top_k=25):
|
| 130 |
+
prob, indices = [], []
|
| 131 |
+
cum_prob = 0.0
|
| 132 |
+
sorted_value, sorted_idx = weighted_scores.softmax(dim=0).sort(
|
| 133 |
+
descending=True, stable=True
|
| 134 |
+
)
|
| 135 |
+
for i in range(len(sorted_idx)):
|
| 136 |
+
# sampling both top-p and numbers.
|
| 137 |
+
if cum_prob < top_p and len(prob) < top_k:
|
| 138 |
+
cum_prob += sorted_value[i]
|
| 139 |
+
prob.append(sorted_value[i])
|
| 140 |
+
indices.append(sorted_idx[i])
|
| 141 |
+
else:
|
| 142 |
+
break
|
| 143 |
+
prob = torch.tensor(prob).to(weighted_scores)
|
| 144 |
+
indices = torch.tensor(indices, dtype=torch.long).to(weighted_scores.device)
|
| 145 |
+
top_ids = indices[prob.multinomial(1, replacement=True)]
|
| 146 |
+
return top_ids
|
| 147 |
+
|
| 148 |
+
|
| 149 |
+
def random_sampling(weighted_scores, decoded_tokens, sampling):
|
| 150 |
+
top_ids = weighted_scores.softmax(dim=0).multinomial(1, replacement=True)
|
| 151 |
+
return top_ids
|
| 152 |
+
|
| 153 |
+
|
| 154 |
+
def fade_in_out(fade_in_mel, fade_out_mel, window):
|
| 155 |
+
device = fade_in_mel.device
|
| 156 |
+
fade_in_mel, fade_out_mel = fade_in_mel.cpu(), fade_out_mel.cpu()
|
| 157 |
+
mel_overlap_len = int(window.shape[0] / 2)
|
| 158 |
+
fade_in_mel[..., :mel_overlap_len] = (
|
| 159 |
+
fade_in_mel[..., :mel_overlap_len] * window[:mel_overlap_len]
|
| 160 |
+
+ fade_out_mel[..., -mel_overlap_len:] * window[mel_overlap_len:]
|
| 161 |
+
)
|
| 162 |
+
return fade_in_mel.to(device)
|
| 163 |
+
|
| 164 |
+
|
| 165 |
+
def set_all_random_seed(seed):
|
| 166 |
+
random.seed(seed)
|
| 167 |
+
np.random.seed(seed)
|
| 168 |
+
torch.manual_seed(seed)
|
| 169 |
+
torch.cuda.manual_seed_all(seed)
|
cosyvoice/utils/executor.py
ADDED
|
@@ -0,0 +1,151 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2020 Mobvoi Inc (Binbin Zhang)
|
| 2 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
|
| 16 |
+
import logging
|
| 17 |
+
from contextlib import nullcontext
|
| 18 |
+
import os
|
| 19 |
+
|
| 20 |
+
import torch
|
| 21 |
+
import torch.distributed as dist
|
| 22 |
+
|
| 23 |
+
from cosyvoice.utils.train_utils import (
|
| 24 |
+
update_parameter_and_lr,
|
| 25 |
+
log_per_step,
|
| 26 |
+
log_per_save,
|
| 27 |
+
batch_forward,
|
| 28 |
+
batch_backward,
|
| 29 |
+
save_model,
|
| 30 |
+
cosyvoice_join,
|
| 31 |
+
)
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
class Executor:
|
| 35 |
+
|
| 36 |
+
def __init__(self):
|
| 37 |
+
self.step = 0
|
| 38 |
+
self.epoch = 0
|
| 39 |
+
self.rank = int(os.environ.get("RANK", 0))
|
| 40 |
+
self.device = torch.device("cuda:{}".format(self.rank))
|
| 41 |
+
|
| 42 |
+
def train_one_epoc(
|
| 43 |
+
self,
|
| 44 |
+
model,
|
| 45 |
+
optimizer,
|
| 46 |
+
scheduler,
|
| 47 |
+
train_data_loader,
|
| 48 |
+
cv_data_loader,
|
| 49 |
+
writer,
|
| 50 |
+
info_dict,
|
| 51 |
+
group_join,
|
| 52 |
+
):
|
| 53 |
+
"""Train one epoch"""
|
| 54 |
+
|
| 55 |
+
lr = optimizer.param_groups[0]["lr"]
|
| 56 |
+
logging.info(
|
| 57 |
+
"Epoch {} TRAIN info lr {} rank {}".format(self.epoch, lr, self.rank)
|
| 58 |
+
)
|
| 59 |
+
logging.info(
|
| 60 |
+
"using accumulate grad, new batch size is {} times"
|
| 61 |
+
" larger than before".format(info_dict["accum_grad"])
|
| 62 |
+
)
|
| 63 |
+
# A context manager to be used in conjunction with an instance of
|
| 64 |
+
# torch.nn.parallel.DistributedDataParallel to be able to train
|
| 65 |
+
# with uneven inputs across participating processes.
|
| 66 |
+
model.train()
|
| 67 |
+
model_context = (
|
| 68 |
+
model.join if info_dict["train_engine"] == "torch_ddp" else nullcontext
|
| 69 |
+
)
|
| 70 |
+
with model_context():
|
| 71 |
+
for batch_idx, batch_dict in enumerate(train_data_loader):
|
| 72 |
+
info_dict["tag"] = "TRAIN"
|
| 73 |
+
info_dict["step"] = self.step
|
| 74 |
+
info_dict["epoch"] = self.epoch
|
| 75 |
+
info_dict["batch_idx"] = batch_idx
|
| 76 |
+
if cosyvoice_join(group_join, info_dict):
|
| 77 |
+
break
|
| 78 |
+
|
| 79 |
+
# Disable gradient synchronizations across DDP processes.
|
| 80 |
+
# Within this context, gradients will be accumulated on module
|
| 81 |
+
# variables, which will later be synchronized.
|
| 82 |
+
if (
|
| 83 |
+
info_dict["train_engine"] == "torch_ddp"
|
| 84 |
+
and (batch_idx + 1) % info_dict["accum_grad"] != 0
|
| 85 |
+
):
|
| 86 |
+
context = model.no_sync
|
| 87 |
+
# Used for single gpu training and DDP gradient synchronization
|
| 88 |
+
# processes.
|
| 89 |
+
else:
|
| 90 |
+
context = nullcontext
|
| 91 |
+
|
| 92 |
+
with context():
|
| 93 |
+
info_dict = batch_forward(model, batch_dict, info_dict)
|
| 94 |
+
info_dict = batch_backward(model, info_dict)
|
| 95 |
+
|
| 96 |
+
info_dict = update_parameter_and_lr(
|
| 97 |
+
model, optimizer, scheduler, info_dict
|
| 98 |
+
)
|
| 99 |
+
log_per_step(writer, info_dict)
|
| 100 |
+
# NOTE specify save_per_step in cosyvoice.yaml if you want to enable step save
|
| 101 |
+
if (
|
| 102 |
+
info_dict["save_per_step"] > 0
|
| 103 |
+
and (self.step + 1) % info_dict["save_per_step"] == 0
|
| 104 |
+
and (batch_idx + 1) % info_dict["accum_grad"] == 0
|
| 105 |
+
):
|
| 106 |
+
dist.barrier()
|
| 107 |
+
self.cv(
|
| 108 |
+
model, cv_data_loader, writer, info_dict, on_batch_end=False
|
| 109 |
+
)
|
| 110 |
+
model.train()
|
| 111 |
+
if (batch_idx + 1) % info_dict["accum_grad"] == 0:
|
| 112 |
+
self.step += 1
|
| 113 |
+
dist.barrier()
|
| 114 |
+
self.cv(model, cv_data_loader, writer, info_dict, on_batch_end=True)
|
| 115 |
+
|
| 116 |
+
@torch.inference_mode()
|
| 117 |
+
def cv(self, model, cv_data_loader, writer, info_dict, on_batch_end=True):
|
| 118 |
+
"""Cross validation on"""
|
| 119 |
+
logging.info(
|
| 120 |
+
"Epoch {} Step {} on_batch_end {} CV rank {}".format(
|
| 121 |
+
self.epoch, self.step + 1, on_batch_end, self.rank
|
| 122 |
+
)
|
| 123 |
+
)
|
| 124 |
+
model.eval()
|
| 125 |
+
total_num_utts, total_loss_dict = 0, {} # avoid division by 0
|
| 126 |
+
for batch_idx, batch_dict in enumerate(cv_data_loader):
|
| 127 |
+
info_dict["tag"] = "CV"
|
| 128 |
+
info_dict["step"] = self.step
|
| 129 |
+
info_dict["epoch"] = self.epoch
|
| 130 |
+
info_dict["batch_idx"] = batch_idx
|
| 131 |
+
|
| 132 |
+
num_utts = len(batch_dict["utts"])
|
| 133 |
+
total_num_utts += num_utts
|
| 134 |
+
|
| 135 |
+
info_dict = batch_forward(model, batch_dict, info_dict)
|
| 136 |
+
|
| 137 |
+
for k, v in info_dict["loss_dict"].items():
|
| 138 |
+
if k not in total_loss_dict:
|
| 139 |
+
total_loss_dict[k] = []
|
| 140 |
+
total_loss_dict[k].append(v.item() * num_utts)
|
| 141 |
+
log_per_step(None, info_dict)
|
| 142 |
+
for k, v in total_loss_dict.items():
|
| 143 |
+
total_loss_dict[k] = sum(v) / total_num_utts
|
| 144 |
+
info_dict["loss_dict"] = total_loss_dict
|
| 145 |
+
log_per_save(writer, info_dict)
|
| 146 |
+
model_name = (
|
| 147 |
+
"epoch_{}_whole".format(self.epoch)
|
| 148 |
+
if on_batch_end
|
| 149 |
+
else "epoch_{}_step_{}".format(self.epoch, self.step + 1)
|
| 150 |
+
)
|
| 151 |
+
save_model(model, model_name, info_dict)
|
cosyvoice/utils/file_utils.py
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2021 Mobvoi Inc. (authors: Binbin Zhang)
|
| 2 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 3 |
+
#
|
| 4 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 5 |
+
# you may not use this file except in compliance with the License.
|
| 6 |
+
# You may obtain a copy of the License at
|
| 7 |
+
#
|
| 8 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 9 |
+
#
|
| 10 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 11 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 12 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 13 |
+
# See the License for the specific language governing permissions and
|
| 14 |
+
# limitations under the License.
|
| 15 |
+
|
| 16 |
+
import json
|
| 17 |
+
import torchaudio
|
| 18 |
+
import logging
|
| 19 |
+
|
| 20 |
+
logging.getLogger("matplotlib").setLevel(logging.WARNING)
|
| 21 |
+
logging.basicConfig(level=logging.DEBUG, format="%(asctime)s %(levelname)s %(message)s")
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
def read_lists(list_file):
|
| 25 |
+
lists = []
|
| 26 |
+
with open(list_file, "r", encoding="utf8") as fin:
|
| 27 |
+
for line in fin:
|
| 28 |
+
lists.append(line.strip())
|
| 29 |
+
return lists
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
def read_json_lists(list_file):
|
| 33 |
+
lists = read_lists(list_file)
|
| 34 |
+
results = {}
|
| 35 |
+
for fn in lists:
|
| 36 |
+
with open(fn, "r", encoding="utf8") as fin:
|
| 37 |
+
results.update(json.load(fin))
|
| 38 |
+
return results
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
def load_wav(wav, target_sr):
|
| 42 |
+
speech, sample_rate = torchaudio.load(wav)
|
| 43 |
+
speech = speech.mean(dim=0, keepdim=True)
|
| 44 |
+
if sample_rate != target_sr:
|
| 45 |
+
# assert sample_rate > target_sr, 'wav sample rate {} must be greater than {}'.format(sample_rate, target_sr)
|
| 46 |
+
speech = torchaudio.transforms.Resample(
|
| 47 |
+
orig_freq=sample_rate, new_freq=target_sr
|
| 48 |
+
)(speech)
|
| 49 |
+
return speech
|
cosyvoice/utils/frontend_utils.py
ADDED
|
@@ -0,0 +1,142 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2024 Alibaba Inc (authors: Xiang Lyu, Zhihao Du)
|
| 2 |
+
#
|
| 3 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 4 |
+
# you may not use this file except in compliance with the License.
|
| 5 |
+
# You may obtain a copy of the License at
|
| 6 |
+
#
|
| 7 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 8 |
+
#
|
| 9 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 10 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 11 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 12 |
+
# See the License for the specific language governing permissions and
|
| 13 |
+
# limitations under the License.
|
| 14 |
+
|
| 15 |
+
import re
|
| 16 |
+
|
| 17 |
+
chinese_char_pattern = re.compile(r"[\u4e00-\u9fff]+")
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
# whether contain chinese character
|
| 21 |
+
def contains_chinese(text):
|
| 22 |
+
return bool(chinese_char_pattern.search(text))
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
# replace special symbol
|
| 26 |
+
def replace_corner_mark(text):
|
| 27 |
+
text = text.replace("²", "平方")
|
| 28 |
+
text = text.replace("³", "立方")
|
| 29 |
+
return text
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
# remove meaningless symbol
|
| 33 |
+
def remove_bracket(text):
|
| 34 |
+
text = text.replace("(", "").replace(")", "")
|
| 35 |
+
text = text.replace("【", "").replace("】", "")
|
| 36 |
+
text = text.replace("`", "").replace("`", "")
|
| 37 |
+
text = text.replace("——", " ")
|
| 38 |
+
return text
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
# spell Arabic numerals
|
| 42 |
+
def spell_out_number(text: str, inflect_parser):
|
| 43 |
+
new_text = []
|
| 44 |
+
st = None
|
| 45 |
+
for i, c in enumerate(text):
|
| 46 |
+
if not c.isdigit():
|
| 47 |
+
if st is not None:
|
| 48 |
+
num_str = inflect_parser.number_to_words(text[st:i])
|
| 49 |
+
new_text.append(num_str)
|
| 50 |
+
st = None
|
| 51 |
+
new_text.append(c)
|
| 52 |
+
else:
|
| 53 |
+
if st is None:
|
| 54 |
+
st = i
|
| 55 |
+
if st is not None and st < len(text):
|
| 56 |
+
num_str = inflect_parser.number_to_words(text[st:])
|
| 57 |
+
new_text.append(num_str)
|
| 58 |
+
return "".join(new_text)
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
# split paragrah logic:
|
| 62 |
+
# 1. per sentence max len token_max_n, min len token_min_n, merge if last sentence len less than merge_len
|
| 63 |
+
# 2. cal sentence len according to lang
|
| 64 |
+
# 3. split sentence according to puncatation
|
| 65 |
+
def split_paragraph(
|
| 66 |
+
text: str,
|
| 67 |
+
tokenize,
|
| 68 |
+
lang="zh",
|
| 69 |
+
token_max_n=80,
|
| 70 |
+
token_min_n=60,
|
| 71 |
+
merge_len=20,
|
| 72 |
+
comma_split=False,
|
| 73 |
+
):
|
| 74 |
+
def calc_utt_length(_text: str):
|
| 75 |
+
if lang == "zh":
|
| 76 |
+
return len(_text)
|
| 77 |
+
else:
|
| 78 |
+
return len(tokenize(_text))
|
| 79 |
+
|
| 80 |
+
def should_merge(_text: str):
|
| 81 |
+
if lang == "zh":
|
| 82 |
+
return len(_text) < merge_len
|
| 83 |
+
else:
|
| 84 |
+
return len(tokenize(_text)) < merge_len
|
| 85 |
+
|
| 86 |
+
if lang == "zh":
|
| 87 |
+
pounc = ["。", "?", "!", ";", ":", "、", ".", "?", "!", ";"]
|
| 88 |
+
else:
|
| 89 |
+
pounc = [".", "?", "!", ";", ":"]
|
| 90 |
+
if comma_split:
|
| 91 |
+
pounc.extend([",", ","])
|
| 92 |
+
|
| 93 |
+
if text[-1] not in pounc:
|
| 94 |
+
if lang == "zh":
|
| 95 |
+
text += "。"
|
| 96 |
+
else:
|
| 97 |
+
text += "."
|
| 98 |
+
|
| 99 |
+
st = 0
|
| 100 |
+
utts = []
|
| 101 |
+
for i, c in enumerate(text):
|
| 102 |
+
if c in pounc:
|
| 103 |
+
if len(text[st:i]) > 0:
|
| 104 |
+
utts.append(text[st:i] + c)
|
| 105 |
+
if i + 1 < len(text) and text[i + 1] in ['"', "”"]:
|
| 106 |
+
tmp = utts.pop(-1)
|
| 107 |
+
utts.append(tmp + text[i + 1])
|
| 108 |
+
st = i + 2
|
| 109 |
+
else:
|
| 110 |
+
st = i + 1
|
| 111 |
+
|
| 112 |
+
final_utts = []
|
| 113 |
+
cur_utt = ""
|
| 114 |
+
for utt in utts:
|
| 115 |
+
if (
|
| 116 |
+
calc_utt_length(cur_utt + utt) > token_max_n
|
| 117 |
+
and calc_utt_length(cur_utt) > token_min_n
|
| 118 |
+
):
|
| 119 |
+
final_utts.append(cur_utt)
|
| 120 |
+
cur_utt = ""
|
| 121 |
+
cur_utt = cur_utt + utt
|
| 122 |
+
if len(cur_utt) > 0:
|
| 123 |
+
if should_merge(cur_utt) and len(final_utts) != 0:
|
| 124 |
+
final_utts[-1] = final_utts[-1] + cur_utt
|
| 125 |
+
else:
|
| 126 |
+
final_utts.append(cur_utt)
|
| 127 |
+
|
| 128 |
+
return final_utts
|
| 129 |
+
|
| 130 |
+
|
| 131 |
+
# remove blank between chinese character
|
| 132 |
+
def replace_blank(text: str):
|
| 133 |
+
out_str = []
|
| 134 |
+
for i, c in enumerate(text):
|
| 135 |
+
if c == " ":
|
| 136 |
+
if (text[i + 1].isascii() and text[i + 1] != " ") and (
|
| 137 |
+
text[i - 1].isascii() and text[i - 1] != " "
|
| 138 |
+
):
|
| 139 |
+
out_str.append(c)
|
| 140 |
+
else:
|
| 141 |
+
out_str.append(c)
|
| 142 |
+
return "".join(out_str)
|
cosyvoice/utils/mask.py
ADDED
|
@@ -0,0 +1,226 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2019 Shigeki Karita
|
| 2 |
+
# 2020 Mobvoi Inc (Binbin Zhang)
|
| 3 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 4 |
+
#
|
| 5 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 6 |
+
# you may not use this file except in compliance with the License.
|
| 7 |
+
# You may obtain a copy of the License at
|
| 8 |
+
#
|
| 9 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 10 |
+
#
|
| 11 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 12 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 13 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 14 |
+
# See the License for the specific language governing permissions and
|
| 15 |
+
# limitations under the License.
|
| 16 |
+
|
| 17 |
+
import torch
|
| 18 |
+
|
| 19 |
+
'''
|
| 20 |
+
def subsequent_mask(
|
| 21 |
+
size: int,
|
| 22 |
+
device: torch.device = torch.device("cpu"),
|
| 23 |
+
) -> torch.Tensor:
|
| 24 |
+
"""Create mask for subsequent steps (size, size).
|
| 25 |
+
|
| 26 |
+
This mask is used only in decoder which works in an auto-regressive mode.
|
| 27 |
+
This means the current step could only do attention with its left steps.
|
| 28 |
+
|
| 29 |
+
In encoder, fully attention is used when streaming is not necessary and
|
| 30 |
+
the sequence is not long. In this case, no attention mask is needed.
|
| 31 |
+
|
| 32 |
+
When streaming is need, chunk-based attention is used in encoder. See
|
| 33 |
+
subsequent_chunk_mask for the chunk-based attention mask.
|
| 34 |
+
|
| 35 |
+
Args:
|
| 36 |
+
size (int): size of mask
|
| 37 |
+
str device (str): "cpu" or "cuda" or torch.Tensor.device
|
| 38 |
+
dtype (torch.device): result dtype
|
| 39 |
+
|
| 40 |
+
Returns:
|
| 41 |
+
torch.Tensor: mask
|
| 42 |
+
|
| 43 |
+
Examples:
|
| 44 |
+
>>> subsequent_mask(3)
|
| 45 |
+
[[1, 0, 0],
|
| 46 |
+
[1, 1, 0],
|
| 47 |
+
[1, 1, 1]]
|
| 48 |
+
"""
|
| 49 |
+
ret = torch.ones(size, size, device=device, dtype=torch.bool)
|
| 50 |
+
return torch.tril(ret)
|
| 51 |
+
'''
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
def subsequent_mask(
|
| 55 |
+
size: int,
|
| 56 |
+
device: torch.device = torch.device("cpu"),
|
| 57 |
+
) -> torch.Tensor:
|
| 58 |
+
"""Create mask for subsequent steps (size, size).
|
| 59 |
+
|
| 60 |
+
This mask is used only in decoder which works in an auto-regressive mode.
|
| 61 |
+
This means the current step could only do attention with its left steps.
|
| 62 |
+
|
| 63 |
+
In encoder, fully attention is used when streaming is not necessary and
|
| 64 |
+
the sequence is not long. In this case, no attention mask is needed.
|
| 65 |
+
|
| 66 |
+
When streaming is need, chunk-based attention is used in encoder. See
|
| 67 |
+
subsequent_chunk_mask for the chunk-based attention mask.
|
| 68 |
+
|
| 69 |
+
Args:
|
| 70 |
+
size (int): size of mask
|
| 71 |
+
str device (str): "cpu" or "cuda" or torch.Tensor.device
|
| 72 |
+
dtype (torch.device): result dtype
|
| 73 |
+
|
| 74 |
+
Returns:
|
| 75 |
+
torch.Tensor: mask
|
| 76 |
+
|
| 77 |
+
Examples:
|
| 78 |
+
>>> subsequent_mask(3)
|
| 79 |
+
[[1, 0, 0],
|
| 80 |
+
[1, 1, 0],
|
| 81 |
+
[1, 1, 1]]
|
| 82 |
+
"""
|
| 83 |
+
arange = torch.arange(size, device=device)
|
| 84 |
+
mask = arange.expand(size, size)
|
| 85 |
+
arange = arange.unsqueeze(-1)
|
| 86 |
+
mask = mask <= arange
|
| 87 |
+
return mask
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
def subsequent_chunk_mask(
|
| 91 |
+
size: int,
|
| 92 |
+
chunk_size: int,
|
| 93 |
+
num_left_chunks: int = -1,
|
| 94 |
+
device: torch.device = torch.device("cpu"),
|
| 95 |
+
) -> torch.Tensor:
|
| 96 |
+
"""Create mask for subsequent steps (size, size) with chunk size,
|
| 97 |
+
this is for streaming encoder
|
| 98 |
+
|
| 99 |
+
Args:
|
| 100 |
+
size (int): size of mask
|
| 101 |
+
chunk_size (int): size of chunk
|
| 102 |
+
num_left_chunks (int): number of left chunks
|
| 103 |
+
<0: use full chunk
|
| 104 |
+
>=0: use num_left_chunks
|
| 105 |
+
device (torch.device): "cpu" or "cuda" or torch.Tensor.device
|
| 106 |
+
|
| 107 |
+
Returns:
|
| 108 |
+
torch.Tensor: mask
|
| 109 |
+
|
| 110 |
+
Examples:
|
| 111 |
+
>>> subsequent_chunk_mask(4, 2)
|
| 112 |
+
[[1, 1, 0, 0],
|
| 113 |
+
[1, 1, 0, 0],
|
| 114 |
+
[1, 1, 1, 1],
|
| 115 |
+
[1, 1, 1, 1]]
|
| 116 |
+
"""
|
| 117 |
+
ret = torch.zeros(size, size, device=device, dtype=torch.bool)
|
| 118 |
+
for i in range(size):
|
| 119 |
+
if num_left_chunks < 0:
|
| 120 |
+
start = 0
|
| 121 |
+
else:
|
| 122 |
+
start = max((i // chunk_size - num_left_chunks) * chunk_size, 0)
|
| 123 |
+
ending = min((i // chunk_size + 1) * chunk_size, size)
|
| 124 |
+
ret[i, start:ending] = True
|
| 125 |
+
return ret
|
| 126 |
+
|
| 127 |
+
|
| 128 |
+
def add_optional_chunk_mask(
|
| 129 |
+
xs: torch.Tensor,
|
| 130 |
+
masks: torch.Tensor,
|
| 131 |
+
use_dynamic_chunk: bool,
|
| 132 |
+
use_dynamic_left_chunk: bool,
|
| 133 |
+
decoding_chunk_size: int,
|
| 134 |
+
static_chunk_size: int,
|
| 135 |
+
num_decoding_left_chunks: int,
|
| 136 |
+
enable_full_context: bool = True,
|
| 137 |
+
):
|
| 138 |
+
"""Apply optional mask for encoder.
|
| 139 |
+
|
| 140 |
+
Args:
|
| 141 |
+
xs (torch.Tensor): padded input, (B, L, D), L for max length
|
| 142 |
+
mask (torch.Tensor): mask for xs, (B, 1, L)
|
| 143 |
+
use_dynamic_chunk (bool): whether to use dynamic chunk or not
|
| 144 |
+
use_dynamic_left_chunk (bool): whether to use dynamic left chunk for
|
| 145 |
+
training.
|
| 146 |
+
decoding_chunk_size (int): decoding chunk size for dynamic chunk, it's
|
| 147 |
+
0: default for training, use random dynamic chunk.
|
| 148 |
+
<0: for decoding, use full chunk.
|
| 149 |
+
>0: for decoding, use fixed chunk size as set.
|
| 150 |
+
static_chunk_size (int): chunk size for static chunk training/decoding
|
| 151 |
+
if it's greater than 0, if use_dynamic_chunk is true,
|
| 152 |
+
this parameter will be ignored
|
| 153 |
+
num_decoding_left_chunks: number of left chunks, this is for decoding,
|
| 154 |
+
the chunk size is decoding_chunk_size.
|
| 155 |
+
>=0: use num_decoding_left_chunks
|
| 156 |
+
<0: use all left chunks
|
| 157 |
+
enable_full_context (bool):
|
| 158 |
+
True: chunk size is either [1, 25] or full context(max_len)
|
| 159 |
+
False: chunk size ~ U[1, 25]
|
| 160 |
+
|
| 161 |
+
Returns:
|
| 162 |
+
torch.Tensor: chunk mask of the input xs.
|
| 163 |
+
"""
|
| 164 |
+
# Whether to use chunk mask or not
|
| 165 |
+
if use_dynamic_chunk:
|
| 166 |
+
max_len = xs.size(1)
|
| 167 |
+
if decoding_chunk_size < 0:
|
| 168 |
+
chunk_size = max_len
|
| 169 |
+
num_left_chunks = -1
|
| 170 |
+
elif decoding_chunk_size > 0:
|
| 171 |
+
chunk_size = decoding_chunk_size
|
| 172 |
+
num_left_chunks = num_decoding_left_chunks
|
| 173 |
+
else:
|
| 174 |
+
# chunk size is either [1, 25] or full context(max_len).
|
| 175 |
+
# Since we use 4 times subsampling and allow up to 1s(100 frames)
|
| 176 |
+
# delay, the maximum frame is 100 / 4 = 25.
|
| 177 |
+
chunk_size = torch.randint(1, max_len, (1,)).item()
|
| 178 |
+
num_left_chunks = -1
|
| 179 |
+
if chunk_size > max_len // 2 and enable_full_context:
|
| 180 |
+
chunk_size = max_len
|
| 181 |
+
else:
|
| 182 |
+
chunk_size = chunk_size % 25 + 1
|
| 183 |
+
if use_dynamic_left_chunk:
|
| 184 |
+
max_left_chunks = (max_len - 1) // chunk_size
|
| 185 |
+
num_left_chunks = torch.randint(0, max_left_chunks, (1,)).item()
|
| 186 |
+
chunk_masks = subsequent_chunk_mask(
|
| 187 |
+
xs.size(1), chunk_size, num_left_chunks, xs.device
|
| 188 |
+
) # (L, L)
|
| 189 |
+
chunk_masks = chunk_masks.unsqueeze(0) # (1, L, L)
|
| 190 |
+
chunk_masks = masks & chunk_masks # (B, L, L)
|
| 191 |
+
elif static_chunk_size > 0:
|
| 192 |
+
num_left_chunks = num_decoding_left_chunks
|
| 193 |
+
chunk_masks = subsequent_chunk_mask(
|
| 194 |
+
xs.size(1), static_chunk_size, num_left_chunks, xs.device
|
| 195 |
+
) # (L, L)
|
| 196 |
+
chunk_masks = chunk_masks.unsqueeze(0) # (1, L, L)
|
| 197 |
+
chunk_masks = masks & chunk_masks # (B, L, L)
|
| 198 |
+
else:
|
| 199 |
+
chunk_masks = masks
|
| 200 |
+
return chunk_masks
|
| 201 |
+
|
| 202 |
+
|
| 203 |
+
def make_pad_mask(lengths: torch.Tensor, max_len: int = 0) -> torch.Tensor:
|
| 204 |
+
"""Make mask tensor containing indices of padded part.
|
| 205 |
+
|
| 206 |
+
See description of make_non_pad_mask.
|
| 207 |
+
|
| 208 |
+
Args:
|
| 209 |
+
lengths (torch.Tensor): Batch of lengths (B,).
|
| 210 |
+
Returns:
|
| 211 |
+
torch.Tensor: Mask tensor containing indices of padded part.
|
| 212 |
+
|
| 213 |
+
Examples:
|
| 214 |
+
>>> lengths = [5, 3, 2]
|
| 215 |
+
>>> make_pad_mask(lengths)
|
| 216 |
+
masks = [[0, 0, 0, 0 ,0],
|
| 217 |
+
[0, 0, 0, 1, 1],
|
| 218 |
+
[0, 0, 1, 1, 1]]
|
| 219 |
+
"""
|
| 220 |
+
batch_size = lengths.size(0)
|
| 221 |
+
max_len = max_len if max_len > 0 else lengths.max().item()
|
| 222 |
+
seq_range = torch.arange(0, max_len, dtype=torch.int64, device=lengths.device)
|
| 223 |
+
seq_range_expand = seq_range.unsqueeze(0).expand(batch_size, max_len)
|
| 224 |
+
seq_length_expand = lengths.unsqueeze(-1)
|
| 225 |
+
mask = seq_range_expand >= seq_length_expand
|
| 226 |
+
return mask
|
cosyvoice/utils/scheduler.py
ADDED
|
@@ -0,0 +1,761 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2020 Mobvoi Inc (Binbin Zhang)
|
| 2 |
+
# 2022 Ximalaya Inc (Yuguang Yang)
|
| 3 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 4 |
+
#
|
| 5 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 6 |
+
# you may not use this file except in compliance with the License.
|
| 7 |
+
# You may obtain a copy of the License at
|
| 8 |
+
#
|
| 9 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 10 |
+
#
|
| 11 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 12 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 13 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 14 |
+
# See the License for the specific language governing permissions and
|
| 15 |
+
# limitations under the License.
|
| 16 |
+
# Modified from ESPnet(https://github.com/espnet/espnet)
|
| 17 |
+
# NeMo(https://github.com/NVIDIA/NeMo)
|
| 18 |
+
|
| 19 |
+
from typing import Union
|
| 20 |
+
|
| 21 |
+
import math
|
| 22 |
+
import warnings
|
| 23 |
+
import torch
|
| 24 |
+
from torch.optim.lr_scheduler import _LRScheduler
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
class WarmupLR(_LRScheduler):
|
| 28 |
+
"""The WarmupLR scheduler
|
| 29 |
+
|
| 30 |
+
This scheduler is almost same as NoamLR Scheduler except for following
|
| 31 |
+
difference:
|
| 32 |
+
|
| 33 |
+
NoamLR:
|
| 34 |
+
lr = optimizer.lr * model_size ** -0.5
|
| 35 |
+
* min(step ** -0.5, step * warmup_step ** -1.5)
|
| 36 |
+
WarmupLR:
|
| 37 |
+
lr = optimizer.lr * warmup_step ** 0.5
|
| 38 |
+
* min(step ** -0.5, step * warmup_step ** -1.5)
|
| 39 |
+
|
| 40 |
+
Note that the maximum lr equals to optimizer.lr in this scheduler.
|
| 41 |
+
|
| 42 |
+
"""
|
| 43 |
+
|
| 44 |
+
def __init__(
|
| 45 |
+
self,
|
| 46 |
+
optimizer: torch.optim.Optimizer,
|
| 47 |
+
warmup_steps: Union[int, float] = 25000,
|
| 48 |
+
last_epoch: int = -1,
|
| 49 |
+
):
|
| 50 |
+
self.warmup_steps = warmup_steps
|
| 51 |
+
|
| 52 |
+
# __init__() must be invoked before setting field
|
| 53 |
+
# because step() is also invoked in __init__()
|
| 54 |
+
super().__init__(optimizer, last_epoch)
|
| 55 |
+
|
| 56 |
+
def __repr__(self):
|
| 57 |
+
return f"{self.__class__.__name__}(warmup_steps={self.warmup_steps})"
|
| 58 |
+
|
| 59 |
+
def get_lr(self):
|
| 60 |
+
step_num = self.last_epoch + 1
|
| 61 |
+
if self.warmup_steps == 0:
|
| 62 |
+
return [lr * step_num**-0.5 for lr in self.base_lrs]
|
| 63 |
+
else:
|
| 64 |
+
return [
|
| 65 |
+
lr
|
| 66 |
+
* self.warmup_steps**0.5
|
| 67 |
+
* min(step_num**-0.5, step_num * self.warmup_steps**-1.5)
|
| 68 |
+
for lr in self.base_lrs
|
| 69 |
+
]
|
| 70 |
+
|
| 71 |
+
def set_step(self, step: int):
|
| 72 |
+
self.last_epoch = step
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
class WarmupPolicy(_LRScheduler):
|
| 76 |
+
"""Adds warmup kwargs and warmup logic to lr policy.
|
| 77 |
+
All arguments should be passed as kwargs for clarity,
|
| 78 |
+
Args:
|
| 79 |
+
warmup_steps: Number of training steps in warmup stage
|
| 80 |
+
warmup_ratio: Ratio of warmup steps to total steps
|
| 81 |
+
max_steps: Total number of steps while training or `None` for
|
| 82 |
+
infinite training
|
| 83 |
+
"""
|
| 84 |
+
|
| 85 |
+
def __init__(
|
| 86 |
+
self,
|
| 87 |
+
optimizer,
|
| 88 |
+
*,
|
| 89 |
+
warmup_steps=None,
|
| 90 |
+
warmup_ratio=None,
|
| 91 |
+
max_steps=None,
|
| 92 |
+
min_lr=0.0,
|
| 93 |
+
last_epoch=-1,
|
| 94 |
+
):
|
| 95 |
+
assert not (
|
| 96 |
+
warmup_steps is not None and warmup_ratio is not None
|
| 97 |
+
), "Either use particular number of step or ratio"
|
| 98 |
+
assert (
|
| 99 |
+
warmup_ratio is None or max_steps is not None
|
| 100 |
+
), "If there is a ratio, there should be a total steps"
|
| 101 |
+
|
| 102 |
+
# It is necessary to assign all attributes *before* __init__,
|
| 103 |
+
# as class is wrapped by an inner class.
|
| 104 |
+
self.max_steps = max_steps
|
| 105 |
+
if warmup_steps is not None:
|
| 106 |
+
self.warmup_steps = warmup_steps
|
| 107 |
+
elif warmup_ratio is not None:
|
| 108 |
+
self.warmup_steps = int(warmup_ratio * max_steps)
|
| 109 |
+
else:
|
| 110 |
+
self.warmup_steps = 0
|
| 111 |
+
|
| 112 |
+
self.min_lr = min_lr
|
| 113 |
+
super().__init__(optimizer, last_epoch)
|
| 114 |
+
|
| 115 |
+
def get_lr(self):
|
| 116 |
+
if not self._get_lr_called_within_step:
|
| 117 |
+
warnings.warn(
|
| 118 |
+
"To get the last learning rate computed "
|
| 119 |
+
"by the scheduler, please use `get_last_lr()`.",
|
| 120 |
+
UserWarning,
|
| 121 |
+
stacklevel=2,
|
| 122 |
+
)
|
| 123 |
+
|
| 124 |
+
step = self.last_epoch
|
| 125 |
+
|
| 126 |
+
if step <= self.warmup_steps and self.warmup_steps > 0:
|
| 127 |
+
return self._get_warmup_lr(step)
|
| 128 |
+
|
| 129 |
+
if step > self.max_steps:
|
| 130 |
+
return [self.min_lr for _ in self.base_lrs]
|
| 131 |
+
|
| 132 |
+
return self._get_lr(step)
|
| 133 |
+
|
| 134 |
+
def _get_warmup_lr(self, step):
|
| 135 |
+
lr_val = (step + 1) / (self.warmup_steps + 1)
|
| 136 |
+
return [initial_lr * lr_val for initial_lr in self.base_lrs]
|
| 137 |
+
|
| 138 |
+
def _get_lr(self, step):
|
| 139 |
+
"""Simple const lr policy"""
|
| 140 |
+
return self.base_lrs
|
| 141 |
+
|
| 142 |
+
|
| 143 |
+
class SquareRootConstantPolicy(_LRScheduler):
|
| 144 |
+
"""Adds warmup kwargs and warmup logic to lr policy.
|
| 145 |
+
All arguments should be passed as kwargs for clarity,
|
| 146 |
+
Args:
|
| 147 |
+
warmup_steps: Number of training steps in warmup stage
|
| 148 |
+
warmup_ratio: Ratio of warmup steps to total steps
|
| 149 |
+
max_steps: Total number of steps while training or `None` for
|
| 150 |
+
infinite training
|
| 151 |
+
"""
|
| 152 |
+
|
| 153 |
+
def __init__(
|
| 154 |
+
self,
|
| 155 |
+
optimizer,
|
| 156 |
+
*,
|
| 157 |
+
constant_steps=None,
|
| 158 |
+
constant_ratio=None,
|
| 159 |
+
max_steps=None,
|
| 160 |
+
min_lr=0.0,
|
| 161 |
+
last_epoch=-1,
|
| 162 |
+
):
|
| 163 |
+
assert not (
|
| 164 |
+
constant_steps is not None and constant_ratio is not None
|
| 165 |
+
), "Either use particular number of step or ratio"
|
| 166 |
+
assert (
|
| 167 |
+
constant_ratio is None or max_steps is not None
|
| 168 |
+
), "If there is a ratio, there should be a total steps"
|
| 169 |
+
|
| 170 |
+
# It is necessary to assign all attributes *before* __init__,
|
| 171 |
+
# as class is wrapped by an inner class.
|
| 172 |
+
self.max_steps = max_steps
|
| 173 |
+
if constant_steps is not None:
|
| 174 |
+
self.constant_steps = constant_steps
|
| 175 |
+
elif constant_ratio is not None:
|
| 176 |
+
self.constant_steps = int(constant_ratio * max_steps)
|
| 177 |
+
else:
|
| 178 |
+
self.constant_steps = 0
|
| 179 |
+
|
| 180 |
+
self.constant_lr = 1 / (constant_steps**0.5)
|
| 181 |
+
self.min_lr = min_lr
|
| 182 |
+
super().__init__(optimizer, last_epoch)
|
| 183 |
+
|
| 184 |
+
def get_lr(self):
|
| 185 |
+
if not self._get_lr_called_within_step:
|
| 186 |
+
warnings.warn(
|
| 187 |
+
"To get the last learning rate computed "
|
| 188 |
+
"by the scheduler, please use `get_last_lr()`.",
|
| 189 |
+
UserWarning,
|
| 190 |
+
stacklevel=2,
|
| 191 |
+
)
|
| 192 |
+
|
| 193 |
+
step = self.last_epoch
|
| 194 |
+
|
| 195 |
+
if step <= self.constant_steps:
|
| 196 |
+
return [self.constant_lr for _ in self.base_lrs]
|
| 197 |
+
|
| 198 |
+
if step > self.max_steps:
|
| 199 |
+
return [self.min_lr for _ in self.base_lrs]
|
| 200 |
+
|
| 201 |
+
return self._get_lr(step)
|
| 202 |
+
|
| 203 |
+
def _get_lr(self, step):
|
| 204 |
+
"""Simple const lr policy"""
|
| 205 |
+
return self.base_lrs
|
| 206 |
+
|
| 207 |
+
|
| 208 |
+
class WarmupHoldPolicy(WarmupPolicy):
|
| 209 |
+
"""Variant of WarmupPolicy which maintains high
|
| 210 |
+
learning rate for a defined number of steps.
|
| 211 |
+
All arguments should be passed as kwargs for clarity,
|
| 212 |
+
Args:
|
| 213 |
+
warmup_steps: Number of training steps in warmup stage
|
| 214 |
+
warmup_ratio: Ratio of warmup steps to total steps
|
| 215 |
+
hold_steps: Number of training steps to
|
| 216 |
+
hold the learning rate after warm up
|
| 217 |
+
hold_ratio: Ratio of hold steps to total steps
|
| 218 |
+
max_steps: Total number of steps while training or `None` for
|
| 219 |
+
infinite training
|
| 220 |
+
"""
|
| 221 |
+
|
| 222 |
+
def __init__(
|
| 223 |
+
self,
|
| 224 |
+
optimizer,
|
| 225 |
+
*,
|
| 226 |
+
warmup_steps=None,
|
| 227 |
+
warmup_ratio=None,
|
| 228 |
+
hold_steps=None,
|
| 229 |
+
hold_ratio=None,
|
| 230 |
+
max_steps=None,
|
| 231 |
+
min_lr=0.0,
|
| 232 |
+
last_epoch=-1,
|
| 233 |
+
):
|
| 234 |
+
assert not (
|
| 235 |
+
hold_steps is not None and hold_ratio is not None
|
| 236 |
+
), "Either use particular number of step or ratio"
|
| 237 |
+
assert (
|
| 238 |
+
hold_ratio is None or max_steps is not None
|
| 239 |
+
), "If there is a ratio, there should be a total steps"
|
| 240 |
+
|
| 241 |
+
self.min_lr = min_lr
|
| 242 |
+
self._last_warmup_lr = 0.0
|
| 243 |
+
|
| 244 |
+
# Necessary to duplicate as class attributes are hidden in inner class
|
| 245 |
+
self.max_steps = max_steps
|
| 246 |
+
if warmup_steps is not None:
|
| 247 |
+
self.warmup_steps = warmup_steps
|
| 248 |
+
elif warmup_ratio is not None:
|
| 249 |
+
self.warmup_steps = int(warmup_ratio * max_steps)
|
| 250 |
+
else:
|
| 251 |
+
self.warmup_steps = 0
|
| 252 |
+
|
| 253 |
+
if hold_steps is not None:
|
| 254 |
+
self.hold_steps = hold_steps + self.warmup_steps
|
| 255 |
+
elif hold_ratio is not None:
|
| 256 |
+
self.hold_steps = int(hold_ratio * max_steps) + self.warmup_steps
|
| 257 |
+
else:
|
| 258 |
+
self.hold_steps = 0
|
| 259 |
+
|
| 260 |
+
super().__init__(
|
| 261 |
+
optimizer,
|
| 262 |
+
warmup_steps=warmup_steps,
|
| 263 |
+
warmup_ratio=warmup_ratio,
|
| 264 |
+
max_steps=max_steps,
|
| 265 |
+
last_epoch=last_epoch,
|
| 266 |
+
min_lr=min_lr,
|
| 267 |
+
)
|
| 268 |
+
|
| 269 |
+
def get_lr(self):
|
| 270 |
+
if not self._get_lr_called_within_step:
|
| 271 |
+
warnings.warn(
|
| 272 |
+
"To get the last learning rate computed by the scheduler,"
|
| 273 |
+
" "
|
| 274 |
+
"please use `get_last_lr()`.",
|
| 275 |
+
UserWarning,
|
| 276 |
+
stacklevel=2,
|
| 277 |
+
)
|
| 278 |
+
|
| 279 |
+
step = self.last_epoch
|
| 280 |
+
|
| 281 |
+
# Warmup phase
|
| 282 |
+
if step <= self.warmup_steps and self.warmup_steps > 0:
|
| 283 |
+
return self._get_warmup_lr(step)
|
| 284 |
+
|
| 285 |
+
# Hold phase
|
| 286 |
+
if (step >= self.warmup_steps) and (step < self.hold_steps):
|
| 287 |
+
return self.base_lrs
|
| 288 |
+
|
| 289 |
+
if step > self.max_steps:
|
| 290 |
+
return [self.min_lr for _ in self.base_lrs]
|
| 291 |
+
|
| 292 |
+
return self._get_lr(step)
|
| 293 |
+
|
| 294 |
+
|
| 295 |
+
class WarmupAnnealHoldPolicy(_LRScheduler):
|
| 296 |
+
"""Adds warmup kwargs and warmup logic to lr policy.
|
| 297 |
+
All arguments should be passed as kwargs for clarity,
|
| 298 |
+
Args:
|
| 299 |
+
warmup_steps: Number of training steps in warmup stage
|
| 300 |
+
warmup_ratio: Ratio of warmup steps to total steps
|
| 301 |
+
max_steps: Total number of steps while training or `None` for
|
| 302 |
+
infinite training
|
| 303 |
+
min_lr: Minimum lr to hold the learning rate after decay at.
|
| 304 |
+
constant_steps: Number of steps to keep lr constant at.
|
| 305 |
+
constant_ratio: Ratio of steps to keep lr constant.
|
| 306 |
+
"""
|
| 307 |
+
|
| 308 |
+
def __init__(
|
| 309 |
+
self,
|
| 310 |
+
optimizer,
|
| 311 |
+
*,
|
| 312 |
+
warmup_steps=None,
|
| 313 |
+
warmup_ratio=None,
|
| 314 |
+
constant_steps=None,
|
| 315 |
+
constant_ratio=None,
|
| 316 |
+
max_steps=None,
|
| 317 |
+
min_lr=0.0,
|
| 318 |
+
last_epoch=-1,
|
| 319 |
+
):
|
| 320 |
+
assert not (
|
| 321 |
+
warmup_steps is not None and warmup_ratio is not None
|
| 322 |
+
), "Either use particular number of step or ratio"
|
| 323 |
+
assert not (
|
| 324 |
+
constant_steps is not None and constant_ratio is not None
|
| 325 |
+
), "Either use constant_steps or constant_ratio"
|
| 326 |
+
assert (
|
| 327 |
+
warmup_ratio is None or max_steps is not None
|
| 328 |
+
), "If there is a ratio, there should be a total steps"
|
| 329 |
+
|
| 330 |
+
# It is necessary to assign all attributes *before* __init__,
|
| 331 |
+
# as class is wrapped by an inner class.
|
| 332 |
+
self.max_steps = max_steps
|
| 333 |
+
|
| 334 |
+
if warmup_steps is not None:
|
| 335 |
+
self.warmup_steps = warmup_steps
|
| 336 |
+
elif warmup_ratio is not None:
|
| 337 |
+
self.warmup_steps = int(warmup_ratio * max_steps)
|
| 338 |
+
else:
|
| 339 |
+
self.warmup_steps = 0
|
| 340 |
+
|
| 341 |
+
if constant_steps is not None:
|
| 342 |
+
self.constant_steps = constant_steps
|
| 343 |
+
elif constant_ratio is not None:
|
| 344 |
+
self.constant_steps = int(constant_ratio * max_steps)
|
| 345 |
+
else:
|
| 346 |
+
self.constant_steps = 0
|
| 347 |
+
|
| 348 |
+
self.decay_steps = max_steps - (self.constant_steps + self.warmup_steps)
|
| 349 |
+
|
| 350 |
+
self.min_lr = min_lr
|
| 351 |
+
super().__init__(optimizer, last_epoch)
|
| 352 |
+
|
| 353 |
+
def get_lr(self):
|
| 354 |
+
if not self._get_lr_called_within_step:
|
| 355 |
+
warnings.warn(
|
| 356 |
+
"To get the last learning rate computed "
|
| 357 |
+
"by the scheduler, please use `get_last_lr()`.",
|
| 358 |
+
UserWarning,
|
| 359 |
+
stacklevel=2,
|
| 360 |
+
)
|
| 361 |
+
|
| 362 |
+
step = self.last_epoch
|
| 363 |
+
|
| 364 |
+
# Warmup steps
|
| 365 |
+
if self.warmup_steps > 0 and step <= self.warmup_steps:
|
| 366 |
+
return self._get_warmup_lr(step)
|
| 367 |
+
|
| 368 |
+
# Constant steps after warmup and decay
|
| 369 |
+
if (
|
| 370 |
+
self.constant_steps > 0
|
| 371 |
+
and (self.warmup_steps + self.decay_steps) < step <= self.max_steps
|
| 372 |
+
):
|
| 373 |
+
return self._get_constant_lr(step)
|
| 374 |
+
|
| 375 |
+
# Min lr after max steps of updates
|
| 376 |
+
if step > self.max_steps:
|
| 377 |
+
return [self.min_lr for _ in self.base_lrs]
|
| 378 |
+
|
| 379 |
+
return self._get_lr(step)
|
| 380 |
+
|
| 381 |
+
def _get_warmup_lr(self, step):
|
| 382 |
+
lr_val = (step + 1) / (self.warmup_steps + 1)
|
| 383 |
+
return [initial_lr * lr_val for initial_lr in self.base_lrs]
|
| 384 |
+
|
| 385 |
+
def _get_constant_lr(self, step):
|
| 386 |
+
return [self.min_lr for _ in self.base_lrs]
|
| 387 |
+
|
| 388 |
+
def _get_lr(self, step):
|
| 389 |
+
"""Simple const lr policy"""
|
| 390 |
+
return self.base_lrs
|
| 391 |
+
|
| 392 |
+
|
| 393 |
+
def _squareroot_annealing(initial_lr, step, max_steps, min_lr):
|
| 394 |
+
mult = ((max_steps - step) / max_steps) ** 0.5
|
| 395 |
+
out_lr = initial_lr * mult
|
| 396 |
+
out_lr = max(out_lr, min_lr)
|
| 397 |
+
return out_lr
|
| 398 |
+
|
| 399 |
+
|
| 400 |
+
def _square_annealing(initial_lr, step, max_steps, min_lr):
|
| 401 |
+
mult = ((max_steps - step) / max_steps) ** 2
|
| 402 |
+
out_lr = initial_lr * mult
|
| 403 |
+
out_lr = max(out_lr, min_lr)
|
| 404 |
+
return out_lr
|
| 405 |
+
|
| 406 |
+
|
| 407 |
+
def _cosine_annealing(initial_lr, step, max_steps, min_lr):
|
| 408 |
+
mult = 0.5 * (1 + math.cos(math.pi * step / max_steps))
|
| 409 |
+
out_lr = (initial_lr - min_lr) * mult + min_lr
|
| 410 |
+
return out_lr
|
| 411 |
+
|
| 412 |
+
|
| 413 |
+
def _linear_warmup_with_cosine_annealing(
|
| 414 |
+
max_lr, warmup_steps, step, decay_steps, min_lr
|
| 415 |
+
):
|
| 416 |
+
assert max_lr > min_lr
|
| 417 |
+
# Use linear warmup for the initial part.
|
| 418 |
+
if warmup_steps > 0 and step <= warmup_steps:
|
| 419 |
+
return max_lr * float(step) / float(warmup_steps)
|
| 420 |
+
|
| 421 |
+
# For any steps larger than `decay_steps`, use `min_lr`.
|
| 422 |
+
if step > warmup_steps + decay_steps:
|
| 423 |
+
return min_lr
|
| 424 |
+
|
| 425 |
+
# If we are done with the warmup period, use the decay style.
|
| 426 |
+
num_steps_ = step - warmup_steps
|
| 427 |
+
decay_steps_ = decay_steps
|
| 428 |
+
decay_ratio = float(num_steps_) / float(decay_steps_)
|
| 429 |
+
assert decay_ratio >= 0.0
|
| 430 |
+
assert decay_ratio <= 1.0
|
| 431 |
+
delta_lr = max_lr - min_lr
|
| 432 |
+
|
| 433 |
+
coeff = 0.5 * (math.cos(math.pi * decay_ratio) + 1.0)
|
| 434 |
+
|
| 435 |
+
return min_lr + coeff * delta_lr
|
| 436 |
+
|
| 437 |
+
|
| 438 |
+
def _poly_decay(initial_lr, step, decay_steps, power, min_lr, cycle):
|
| 439 |
+
if cycle:
|
| 440 |
+
multiplier = 1.0 if step == 0 else math.ceil(step / decay_steps)
|
| 441 |
+
decay_steps *= multiplier
|
| 442 |
+
else:
|
| 443 |
+
step = min(step, decay_steps)
|
| 444 |
+
p = step / decay_steps
|
| 445 |
+
lr = (initial_lr - min_lr) * math.pow(1.0 - p, power)
|
| 446 |
+
lr += min_lr
|
| 447 |
+
return lr
|
| 448 |
+
|
| 449 |
+
|
| 450 |
+
def _noam_hold_annealing(
|
| 451 |
+
initial_lr, step, warmup_steps, hold_steps, decay_rate, min_lr
|
| 452 |
+
):
|
| 453 |
+
# hold_steps = total number of steps
|
| 454 |
+
# to hold the LR, not the warmup + hold steps.
|
| 455 |
+
T_warmup_decay = max(1, warmup_steps**decay_rate)
|
| 456 |
+
T_hold_decay = max(1, (step - hold_steps) ** decay_rate)
|
| 457 |
+
lr = (initial_lr * T_warmup_decay) / T_hold_decay
|
| 458 |
+
lr = max(lr, min_lr)
|
| 459 |
+
return lr
|
| 460 |
+
|
| 461 |
+
|
| 462 |
+
class SquareAnnealing(WarmupPolicy):
|
| 463 |
+
|
| 464 |
+
def __init__(self, optimizer, *, max_steps, min_lr=1e-5, last_epoch=-1, **kwargs):
|
| 465 |
+
super().__init__(
|
| 466 |
+
optimizer=optimizer,
|
| 467 |
+
max_steps=max_steps,
|
| 468 |
+
last_epoch=last_epoch,
|
| 469 |
+
min_lr=min_lr,
|
| 470 |
+
**kwargs,
|
| 471 |
+
)
|
| 472 |
+
|
| 473 |
+
def _get_lr(self, step):
|
| 474 |
+
new_lrs = [
|
| 475 |
+
_square_annealing(
|
| 476 |
+
initial_lr=initial_lr,
|
| 477 |
+
step=step - self.warmup_steps,
|
| 478 |
+
max_steps=self.max_steps - self.warmup_steps,
|
| 479 |
+
min_lr=self.min_lr,
|
| 480 |
+
)
|
| 481 |
+
for initial_lr in self.base_lrs
|
| 482 |
+
]
|
| 483 |
+
return new_lrs
|
| 484 |
+
|
| 485 |
+
|
| 486 |
+
class SquareRootAnnealing(WarmupPolicy):
|
| 487 |
+
|
| 488 |
+
def __init__(self, optimizer, *, max_steps, min_lr=0, last_epoch=-1, **kwargs):
|
| 489 |
+
super().__init__(
|
| 490 |
+
optimizer=optimizer,
|
| 491 |
+
max_steps=max_steps,
|
| 492 |
+
last_epoch=last_epoch,
|
| 493 |
+
min_lr=min_lr,
|
| 494 |
+
**kwargs,
|
| 495 |
+
)
|
| 496 |
+
|
| 497 |
+
def _get_lr(self, step):
|
| 498 |
+
new_lrs = [
|
| 499 |
+
_squareroot_annealing(
|
| 500 |
+
initial_lr=initial_lr,
|
| 501 |
+
step=step,
|
| 502 |
+
max_steps=self.max_steps,
|
| 503 |
+
min_lr=self.min_lr,
|
| 504 |
+
)
|
| 505 |
+
for initial_lr in self.base_lrs
|
| 506 |
+
]
|
| 507 |
+
return new_lrs
|
| 508 |
+
|
| 509 |
+
|
| 510 |
+
class CosineAnnealing(WarmupAnnealHoldPolicy):
|
| 511 |
+
|
| 512 |
+
def __init__(self, optimizer, *, max_steps, min_lr=0, last_epoch=-1, **kwargs):
|
| 513 |
+
super().__init__(
|
| 514 |
+
optimizer=optimizer,
|
| 515 |
+
max_steps=max_steps,
|
| 516 |
+
last_epoch=last_epoch,
|
| 517 |
+
min_lr=min_lr,
|
| 518 |
+
**kwargs,
|
| 519 |
+
)
|
| 520 |
+
|
| 521 |
+
def _get_lr(self, step):
|
| 522 |
+
for initial_lr in self.base_lrs:
|
| 523 |
+
if initial_lr < self.min_lr:
|
| 524 |
+
raise ValueError(
|
| 525 |
+
f"{self} received an initial learning rate "
|
| 526 |
+
f"that was lower than the minimum learning rate."
|
| 527 |
+
)
|
| 528 |
+
|
| 529 |
+
if self.constant_steps is None or self.constant_steps == 0:
|
| 530 |
+
new_lrs = [
|
| 531 |
+
_cosine_annealing(
|
| 532 |
+
initial_lr=initial_lr,
|
| 533 |
+
step=step - self.warmup_steps,
|
| 534 |
+
max_steps=self.max_steps - self.warmup_steps,
|
| 535 |
+
min_lr=self.min_lr,
|
| 536 |
+
)
|
| 537 |
+
for initial_lr in self.base_lrs
|
| 538 |
+
]
|
| 539 |
+
else:
|
| 540 |
+
new_lrs = self._get_linear_warmup_with_cosine_annealing_lr(step)
|
| 541 |
+
return new_lrs
|
| 542 |
+
|
| 543 |
+
def _get_warmup_lr(self, step):
|
| 544 |
+
if self.constant_steps is None or self.constant_steps == 0:
|
| 545 |
+
return super()._get_warmup_lr(step)
|
| 546 |
+
else:
|
| 547 |
+
# Use linear warmup for the initial part.
|
| 548 |
+
return self._get_linear_warmup_with_cosine_annealing_lr(step)
|
| 549 |
+
|
| 550 |
+
def _get_constant_lr(self, step):
|
| 551 |
+
# Only called when `constant_steps` > 0.
|
| 552 |
+
return self._get_linear_warmup_with_cosine_annealing_lr(step)
|
| 553 |
+
|
| 554 |
+
def _get_linear_warmup_with_cosine_annealing_lr(self, step):
|
| 555 |
+
# Cosine Schedule for Megatron LM,
|
| 556 |
+
# slightly different warmup schedule + constant LR at the end.
|
| 557 |
+
new_lrs = [
|
| 558 |
+
_linear_warmup_with_cosine_annealing(
|
| 559 |
+
max_lr=self.base_lrs[0],
|
| 560 |
+
warmup_steps=self.warmup_steps,
|
| 561 |
+
step=step,
|
| 562 |
+
decay_steps=self.decay_steps,
|
| 563 |
+
min_lr=self.min_lr,
|
| 564 |
+
)
|
| 565 |
+
for _ in self.base_lrs
|
| 566 |
+
]
|
| 567 |
+
return new_lrs
|
| 568 |
+
|
| 569 |
+
|
| 570 |
+
class NoamAnnealing(_LRScheduler):
|
| 571 |
+
|
| 572 |
+
def __init__(
|
| 573 |
+
self,
|
| 574 |
+
optimizer,
|
| 575 |
+
*,
|
| 576 |
+
d_model,
|
| 577 |
+
warmup_steps=None,
|
| 578 |
+
warmup_ratio=None,
|
| 579 |
+
max_steps=None,
|
| 580 |
+
min_lr=0.0,
|
| 581 |
+
last_epoch=-1,
|
| 582 |
+
):
|
| 583 |
+
self._normalize = d_model ** (-0.5)
|
| 584 |
+
assert not (
|
| 585 |
+
warmup_steps is not None and warmup_ratio is not None
|
| 586 |
+
), "Either use particular number of step or ratio"
|
| 587 |
+
assert (
|
| 588 |
+
warmup_ratio is None or max_steps is not None
|
| 589 |
+
), "If there is a ratio, there should be a total steps"
|
| 590 |
+
|
| 591 |
+
# It is necessary to assign all attributes *before* __init__,
|
| 592 |
+
# as class is wrapped by an inner class.
|
| 593 |
+
self.max_steps = max_steps
|
| 594 |
+
if warmup_steps is not None:
|
| 595 |
+
self.warmup_steps = warmup_steps
|
| 596 |
+
elif warmup_ratio is not None:
|
| 597 |
+
self.warmup_steps = int(warmup_ratio * max_steps)
|
| 598 |
+
else:
|
| 599 |
+
self.warmup_steps = 0
|
| 600 |
+
|
| 601 |
+
self.min_lr = min_lr
|
| 602 |
+
super().__init__(optimizer, last_epoch)
|
| 603 |
+
|
| 604 |
+
def get_lr(self):
|
| 605 |
+
if not self._get_lr_called_within_step:
|
| 606 |
+
warnings.warn(
|
| 607 |
+
"To get the last learning rate computed "
|
| 608 |
+
"by the scheduler, please use `get_last_lr()`.",
|
| 609 |
+
UserWarning,
|
| 610 |
+
stacklevel=2,
|
| 611 |
+
)
|
| 612 |
+
|
| 613 |
+
step = max(1, self.last_epoch)
|
| 614 |
+
|
| 615 |
+
for initial_lr in self.base_lrs:
|
| 616 |
+
if initial_lr < self.min_lr:
|
| 617 |
+
raise ValueError(
|
| 618 |
+
f"{self} received an initial learning rate "
|
| 619 |
+
f"that was lower than the minimum learning rate."
|
| 620 |
+
)
|
| 621 |
+
|
| 622 |
+
new_lrs = [
|
| 623 |
+
self._noam_annealing(initial_lr=initial_lr, step=step)
|
| 624 |
+
for initial_lr in self.base_lrs
|
| 625 |
+
]
|
| 626 |
+
return new_lrs
|
| 627 |
+
|
| 628 |
+
def _noam_annealing(self, initial_lr, step):
|
| 629 |
+
if self.warmup_steps > 0:
|
| 630 |
+
mult = self._normalize * min(
|
| 631 |
+
step ** (-0.5), step * (self.warmup_steps ** (-1.5))
|
| 632 |
+
)
|
| 633 |
+
else:
|
| 634 |
+
mult = self._normalize * step ** (-0.5)
|
| 635 |
+
|
| 636 |
+
out_lr = initial_lr * mult
|
| 637 |
+
if step > self.warmup_steps:
|
| 638 |
+
out_lr = max(out_lr, self.min_lr)
|
| 639 |
+
return out_lr
|
| 640 |
+
|
| 641 |
+
|
| 642 |
+
class NoamHoldAnnealing(WarmupHoldPolicy):
|
| 643 |
+
|
| 644 |
+
def __init__(
|
| 645 |
+
self,
|
| 646 |
+
optimizer,
|
| 647 |
+
*,
|
| 648 |
+
max_steps,
|
| 649 |
+
decay_rate=0.5,
|
| 650 |
+
min_lr=0.0,
|
| 651 |
+
last_epoch=-1,
|
| 652 |
+
**kwargs,
|
| 653 |
+
):
|
| 654 |
+
"""
|
| 655 |
+
From Nemo:
|
| 656 |
+
Implementation of the Noam Hold Annealing policy
|
| 657 |
+
from the SqueezeFormer paper.
|
| 658 |
+
|
| 659 |
+
Unlike NoamAnnealing, the peak learning rate
|
| 660 |
+
can be explicitly set for this scheduler.
|
| 661 |
+
The schedule first performs linear warmup,
|
| 662 |
+
then holds the peak LR, then decays with some schedule for
|
| 663 |
+
the remainder of the steps.
|
| 664 |
+
Therefore the min-lr is still dependent
|
| 665 |
+
on the hyper parameters selected.
|
| 666 |
+
|
| 667 |
+
It's schedule is determined by three factors-
|
| 668 |
+
|
| 669 |
+
Warmup Steps: Initial stage, where linear warmup
|
| 670 |
+
occurs uptil the peak LR is reached. Unlike NoamAnnealing,
|
| 671 |
+
the peak LR is explicitly stated here instead of a scaling factor.
|
| 672 |
+
|
| 673 |
+
Hold Steps: Intermediate stage, where the peak LR
|
| 674 |
+
is maintained for some number of steps. In this region,
|
| 675 |
+
the high peak LR allows the model to converge faster
|
| 676 |
+
if training is stable. However the high LR
|
| 677 |
+
may also cause instability during training.
|
| 678 |
+
Should usually be a significant fraction of training
|
| 679 |
+
steps (around 30-40% of the entire training steps).
|
| 680 |
+
|
| 681 |
+
Decay Steps: Final stage, where the LR rapidly decays
|
| 682 |
+
with some scaling rate (set by decay rate).
|
| 683 |
+
To attain Noam decay, use 0.5,
|
| 684 |
+
for Squeezeformer recommended decay, use 1.0.
|
| 685 |
+
The fast decay after prolonged high LR during
|
| 686 |
+
hold phase allows for rapid convergence.
|
| 687 |
+
|
| 688 |
+
References:
|
| 689 |
+
- [Squeezeformer:
|
| 690 |
+
An Efficient Transformer for Automatic Speech Recognition]
|
| 691 |
+
(https://arxiv.org/abs/2206.00888)
|
| 692 |
+
|
| 693 |
+
Args:
|
| 694 |
+
optimizer: Pytorch compatible Optimizer object.
|
| 695 |
+
warmup_steps: Number of training steps in warmup stage
|
| 696 |
+
warmup_ratio: Ratio of warmup steps to total steps
|
| 697 |
+
hold_steps: Number of training steps to
|
| 698 |
+
hold the learning rate after warm up
|
| 699 |
+
hold_ratio: Ratio of hold steps to total steps
|
| 700 |
+
max_steps: Total number of steps while training or `None` for
|
| 701 |
+
infinite training
|
| 702 |
+
decay_rate: Float value describing the polynomial decay
|
| 703 |
+
after the hold period. Default value
|
| 704 |
+
of 0.5 corresponds to Noam decay.
|
| 705 |
+
min_lr: Minimum learning rate.
|
| 706 |
+
"""
|
| 707 |
+
self.decay_rate = decay_rate
|
| 708 |
+
super().__init__(
|
| 709 |
+
optimizer=optimizer,
|
| 710 |
+
max_steps=max_steps,
|
| 711 |
+
last_epoch=last_epoch,
|
| 712 |
+
min_lr=min_lr,
|
| 713 |
+
**kwargs,
|
| 714 |
+
)
|
| 715 |
+
|
| 716 |
+
def _get_lr(self, step):
|
| 717 |
+
if self.warmup_steps is None or self.warmup_steps == 0:
|
| 718 |
+
raise ValueError("Noam scheduler cannot be used without warmup steps")
|
| 719 |
+
|
| 720 |
+
if self.hold_steps > 0:
|
| 721 |
+
hold_steps = self.hold_steps - self.warmup_steps
|
| 722 |
+
else:
|
| 723 |
+
hold_steps = 0
|
| 724 |
+
|
| 725 |
+
new_lrs = [
|
| 726 |
+
_noam_hold_annealing(
|
| 727 |
+
initial_lr,
|
| 728 |
+
step=step,
|
| 729 |
+
warmup_steps=self.warmup_steps,
|
| 730 |
+
hold_steps=hold_steps,
|
| 731 |
+
decay_rate=self.decay_rate,
|
| 732 |
+
min_lr=self.min_lr,
|
| 733 |
+
)
|
| 734 |
+
for initial_lr in self.base_lrs
|
| 735 |
+
]
|
| 736 |
+
return new_lrs
|
| 737 |
+
|
| 738 |
+
def set_step(self, step: int):
|
| 739 |
+
self.last_epoch = step
|
| 740 |
+
|
| 741 |
+
|
| 742 |
+
class ConstantLR(_LRScheduler):
|
| 743 |
+
"""The ConstantLR scheduler
|
| 744 |
+
|
| 745 |
+
This scheduler keeps a constant lr
|
| 746 |
+
|
| 747 |
+
"""
|
| 748 |
+
|
| 749 |
+
def __init__(
|
| 750 |
+
self,
|
| 751 |
+
optimizer: torch.optim.Optimizer,
|
| 752 |
+
):
|
| 753 |
+
# __init__() must be invoked before setting field
|
| 754 |
+
# because step() is also invoked in __init__()
|
| 755 |
+
super().__init__(optimizer)
|
| 756 |
+
|
| 757 |
+
def get_lr(self):
|
| 758 |
+
return self.base_lrs
|
| 759 |
+
|
| 760 |
+
def set_step(self, step: int):
|
| 761 |
+
self.last_epoch = step
|
cosyvoice/utils/train_utils.py
ADDED
|
@@ -0,0 +1,350 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Copyright (c) 2021 Mobvoi Inc. (authors: Binbin Zhang)
|
| 2 |
+
# 2023 Horizon Inc. (authors: Xingchen Song)
|
| 3 |
+
# 2024 Alibaba Inc (authors: Xiang Lyu)
|
| 4 |
+
#
|
| 5 |
+
# Licensed under the Apache License, Version 2.0 (the "License");
|
| 6 |
+
# you may not use this file except in compliance with the License.
|
| 7 |
+
# You may obtain a copy of the License at
|
| 8 |
+
#
|
| 9 |
+
# http://www.apache.org/licenses/LICENSE-2.0
|
| 10 |
+
#
|
| 11 |
+
# Unless required by applicable law or agreed to in writing, software
|
| 12 |
+
# distributed under the License is distributed on an "AS IS" BASIS,
|
| 13 |
+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 14 |
+
# See the License for the specific language governing permissions and
|
| 15 |
+
# limitations under the License.
|
| 16 |
+
|
| 17 |
+
from contextlib import nullcontext
|
| 18 |
+
import logging
|
| 19 |
+
import os
|
| 20 |
+
import torch
|
| 21 |
+
import json
|
| 22 |
+
import re
|
| 23 |
+
import datetime
|
| 24 |
+
import yaml
|
| 25 |
+
|
| 26 |
+
import deepspeed
|
| 27 |
+
import torch.optim as optim
|
| 28 |
+
import torch.distributed as dist
|
| 29 |
+
|
| 30 |
+
from torch.utils.tensorboard import SummaryWriter
|
| 31 |
+
from torch.utils.data import DataLoader
|
| 32 |
+
from torch.nn.utils import clip_grad_norm_
|
| 33 |
+
|
| 34 |
+
from deepspeed.runtime.zero.stage_1_and_2 import (
|
| 35 |
+
estimate_zero2_model_states_mem_needs_all_live,
|
| 36 |
+
)
|
| 37 |
+
|
| 38 |
+
from cosyvoice.dataset.dataset import Dataset
|
| 39 |
+
from cosyvoice.utils.scheduler import (
|
| 40 |
+
WarmupLR,
|
| 41 |
+
NoamHoldAnnealing,
|
| 42 |
+
ConstantLR,
|
| 43 |
+
)
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
def init_distributed(args):
|
| 47 |
+
world_size = int(os.environ.get("WORLD_SIZE", 1))
|
| 48 |
+
local_rank = int(os.environ.get("LOCAL_RANK", 0))
|
| 49 |
+
rank = int(os.environ.get("RANK", 0))
|
| 50 |
+
logging.info(
|
| 51 |
+
"training on multiple gpus, this gpu {}".format(local_rank)
|
| 52 |
+
+ ", rank {}, world_size {}".format(rank, world_size)
|
| 53 |
+
)
|
| 54 |
+
if args.train_engine == "torch_ddp":
|
| 55 |
+
torch.cuda.set_device(local_rank)
|
| 56 |
+
dist.init_process_group(args.dist_backend)
|
| 57 |
+
else:
|
| 58 |
+
deepspeed.init_distributed(dist_backend=args.dist_backend)
|
| 59 |
+
return world_size, local_rank, rank
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
def init_dataset_and_dataloader(args, configs):
|
| 63 |
+
train_dataset = Dataset(
|
| 64 |
+
args.train_data,
|
| 65 |
+
data_pipeline=configs["data_pipeline"],
|
| 66 |
+
mode="train",
|
| 67 |
+
shuffle=True,
|
| 68 |
+
partition=True,
|
| 69 |
+
)
|
| 70 |
+
cv_dataset = Dataset(
|
| 71 |
+
args.cv_data,
|
| 72 |
+
data_pipeline=configs["data_pipeline"],
|
| 73 |
+
mode="train",
|
| 74 |
+
shuffle=False,
|
| 75 |
+
partition=False,
|
| 76 |
+
)
|
| 77 |
+
|
| 78 |
+
# do not use persistent_workers=True, as whisper tokenizer opens tiktoken file each time when the for loop starts
|
| 79 |
+
train_data_loader = DataLoader(
|
| 80 |
+
train_dataset,
|
| 81 |
+
batch_size=None,
|
| 82 |
+
pin_memory=args.pin_memory,
|
| 83 |
+
num_workers=args.num_workers,
|
| 84 |
+
prefetch_factor=args.prefetch,
|
| 85 |
+
)
|
| 86 |
+
cv_data_loader = DataLoader(
|
| 87 |
+
cv_dataset,
|
| 88 |
+
batch_size=None,
|
| 89 |
+
pin_memory=args.pin_memory,
|
| 90 |
+
num_workers=args.num_workers,
|
| 91 |
+
prefetch_factor=args.prefetch,
|
| 92 |
+
)
|
| 93 |
+
return train_dataset, cv_dataset, train_data_loader, cv_data_loader
|
| 94 |
+
|
| 95 |
+
|
| 96 |
+
def check_modify_and_save_config(args, configs):
|
| 97 |
+
if args.train_engine == "torch_ddp":
|
| 98 |
+
configs["train_conf"]["dtype"] = "fp32"
|
| 99 |
+
else:
|
| 100 |
+
with open(args.deepspeed_config, "r") as fin:
|
| 101 |
+
ds_configs = json.load(fin)
|
| 102 |
+
if "fp16" in ds_configs and ds_configs["fp16"]["enabled"]:
|
| 103 |
+
configs["train_conf"]["dtype"] = "fp16"
|
| 104 |
+
elif "bf16" in ds_configs and ds_configs["bf16"]["enabled"]:
|
| 105 |
+
configs["train_conf"]["dtype"] = "bf16"
|
| 106 |
+
else:
|
| 107 |
+
configs["train_conf"]["dtype"] = "fp32"
|
| 108 |
+
assert ds_configs["train_micro_batch_size_per_gpu"] == 1
|
| 109 |
+
# if use deepspeed, override ddp config
|
| 110 |
+
configs["train_conf"]["save_per_step"] = int(
|
| 111 |
+
configs["train_conf"]["save_per_step"]
|
| 112 |
+
* configs["train_conf"]["accum_grad"]
|
| 113 |
+
/ ds_configs["gradient_accumulation_steps"]
|
| 114 |
+
)
|
| 115 |
+
configs["train_conf"]["accum_grad"] = ds_configs["gradient_accumulation_steps"]
|
| 116 |
+
configs["train_conf"]["grad_clip"] = ds_configs["gradient_clipping"]
|
| 117 |
+
configs["train_conf"]["log_interval"] = ds_configs["steps_per_print"]
|
| 118 |
+
return configs
|
| 119 |
+
|
| 120 |
+
|
| 121 |
+
def wrap_cuda_model(args, model):
|
| 122 |
+
local_world_size = int(os.environ.get("LOCAL_WORLD_SIZE", 1))
|
| 123 |
+
world_size = int(os.environ.get("WORLD_SIZE", 1))
|
| 124 |
+
if args.train_engine == "torch_ddp": # native pytorch ddp
|
| 125 |
+
assert torch.cuda.is_available()
|
| 126 |
+
model.cuda()
|
| 127 |
+
model = torch.nn.parallel.DistributedDataParallel(
|
| 128 |
+
model, find_unused_parameters=True
|
| 129 |
+
)
|
| 130 |
+
else:
|
| 131 |
+
if int(os.environ.get("RANK", 0)) == 0:
|
| 132 |
+
logging.info("Estimating model states memory needs (zero2)...")
|
| 133 |
+
estimate_zero2_model_states_mem_needs_all_live(
|
| 134 |
+
model,
|
| 135 |
+
num_gpus_per_node=local_world_size,
|
| 136 |
+
num_nodes=world_size // local_world_size,
|
| 137 |
+
)
|
| 138 |
+
return model
|
| 139 |
+
|
| 140 |
+
|
| 141 |
+
def init_optimizer_and_scheduler(args, configs, model):
|
| 142 |
+
if configs["train_conf"]["optim"] == "adam":
|
| 143 |
+
optimizer = optim.Adam(
|
| 144 |
+
model.parameters(), **configs["train_conf"]["optim_conf"]
|
| 145 |
+
)
|
| 146 |
+
elif configs["train_conf"]["optim"] == "adamw":
|
| 147 |
+
optimizer = optim.AdamW(
|
| 148 |
+
model.parameters(), **configs["train_conf"]["optim_conf"]
|
| 149 |
+
)
|
| 150 |
+
else:
|
| 151 |
+
raise ValueError("unknown optimizer: " + configs["train_conf"])
|
| 152 |
+
|
| 153 |
+
if configs["train_conf"]["scheduler"] == "warmuplr":
|
| 154 |
+
scheduler_type = WarmupLR
|
| 155 |
+
scheduler = WarmupLR(optimizer, **configs["train_conf"]["scheduler_conf"])
|
| 156 |
+
elif configs["train_conf"]["scheduler"] == "NoamHoldAnnealing":
|
| 157 |
+
scheduler_type = NoamHoldAnnealing
|
| 158 |
+
scheduler = NoamHoldAnnealing(
|
| 159 |
+
optimizer, **configs["train_conf"]["scheduler_conf"]
|
| 160 |
+
)
|
| 161 |
+
elif configs["train_conf"]["scheduler"] == "constantlr":
|
| 162 |
+
scheduler_type = ConstantLR
|
| 163 |
+
scheduler = ConstantLR(optimizer)
|
| 164 |
+
else:
|
| 165 |
+
raise ValueError("unknown scheduler: " + configs["train_conf"])
|
| 166 |
+
|
| 167 |
+
# use deepspeed optimizer for speedup
|
| 168 |
+
if args.train_engine == "deepspeed":
|
| 169 |
+
|
| 170 |
+
def scheduler(opt):
|
| 171 |
+
return scheduler_type(opt, **configs["train_conf"]["scheduler_conf"])
|
| 172 |
+
|
| 173 |
+
model, optimizer, _, scheduler = deepspeed.initialize(
|
| 174 |
+
args=args,
|
| 175 |
+
model=model,
|
| 176 |
+
optimizer=None,
|
| 177 |
+
lr_scheduler=scheduler,
|
| 178 |
+
model_parameters=model.parameters(),
|
| 179 |
+
)
|
| 180 |
+
|
| 181 |
+
return model, optimizer, scheduler
|
| 182 |
+
|
| 183 |
+
|
| 184 |
+
def init_summarywriter(args):
|
| 185 |
+
writer = None
|
| 186 |
+
if int(os.environ.get("RANK", 0)) == 0:
|
| 187 |
+
os.makedirs(args.model_dir, exist_ok=True)
|
| 188 |
+
writer = SummaryWriter(args.tensorboard_dir)
|
| 189 |
+
return writer
|
| 190 |
+
|
| 191 |
+
|
| 192 |
+
def save_model(model, model_name, info_dict):
|
| 193 |
+
rank = int(os.environ.get("RANK", 0))
|
| 194 |
+
model_dir = info_dict["model_dir"]
|
| 195 |
+
save_model_path = os.path.join(model_dir, "{}.pt".format(model_name))
|
| 196 |
+
|
| 197 |
+
if info_dict["train_engine"] == "torch_ddp":
|
| 198 |
+
if rank == 0:
|
| 199 |
+
torch.save(model.module.state_dict(), save_model_path)
|
| 200 |
+
else:
|
| 201 |
+
with torch.no_grad():
|
| 202 |
+
model.save_checkpoint(
|
| 203 |
+
save_dir=model_dir, tag=model_name, client_state=info_dict
|
| 204 |
+
)
|
| 205 |
+
if rank == 0:
|
| 206 |
+
info_path = re.sub(".pt$", ".yaml", save_model_path)
|
| 207 |
+
info_dict["save_time"] = datetime.datetime.now().strftime("%d/%m/%Y %H:%M:%S")
|
| 208 |
+
with open(info_path, "w") as fout:
|
| 209 |
+
data = yaml.dump(info_dict)
|
| 210 |
+
fout.write(data)
|
| 211 |
+
logging.info(
|
| 212 |
+
"[Rank {}] Checkpoint: save to checkpoint {}".format(rank, save_model_path)
|
| 213 |
+
)
|
| 214 |
+
|
| 215 |
+
|
| 216 |
+
def cosyvoice_join(group_join, info_dict):
|
| 217 |
+
world_size = int(os.environ.get("WORLD_SIZE", 1))
|
| 218 |
+
local_rank = int(os.environ.get("LOCAL_RANK", 0))
|
| 219 |
+
rank = int(os.environ.get("RANK", 0))
|
| 220 |
+
|
| 221 |
+
if info_dict["batch_idx"] != 0:
|
| 222 |
+
# we try to join all rank in both ddp and deepspeed mode, in case different rank has different lr
|
| 223 |
+
try:
|
| 224 |
+
dist.monitored_barrier(
|
| 225 |
+
group=group_join, timeout=group_join.options._timeout
|
| 226 |
+
)
|
| 227 |
+
return False
|
| 228 |
+
except RuntimeError as e:
|
| 229 |
+
logging.info(
|
| 230 |
+
"Detected uneven workload distribution: {}\n".format(e)
|
| 231 |
+
+ "Break current worker to manually join all workers, "
|
| 232 |
+
+ "world_size {}, current rank {}, current local_rank {}\n".format(
|
| 233 |
+
world_size, rank, local_rank
|
| 234 |
+
)
|
| 235 |
+
)
|
| 236 |
+
return True
|
| 237 |
+
else:
|
| 238 |
+
return False
|
| 239 |
+
|
| 240 |
+
|
| 241 |
+
def batch_forward(model, batch, info_dict):
|
| 242 |
+
device = int(os.environ.get("LOCAL_RANK", 0))
|
| 243 |
+
|
| 244 |
+
dtype = info_dict["dtype"]
|
| 245 |
+
if dtype == "fp16":
|
| 246 |
+
dtype = torch.float16
|
| 247 |
+
elif dtype == "bf16":
|
| 248 |
+
dtype = torch.bfloat16
|
| 249 |
+
else: # fp32
|
| 250 |
+
dtype = torch.float32
|
| 251 |
+
|
| 252 |
+
if info_dict["train_engine"] == "torch_ddp":
|
| 253 |
+
autocast = nullcontext()
|
| 254 |
+
else:
|
| 255 |
+
autocast = torch.cuda.amp.autocast(
|
| 256 |
+
enabled=True, dtype=dtype, cache_enabled=False
|
| 257 |
+
)
|
| 258 |
+
|
| 259 |
+
with autocast:
|
| 260 |
+
info_dict["loss_dict"] = model(batch, device)
|
| 261 |
+
return info_dict
|
| 262 |
+
|
| 263 |
+
|
| 264 |
+
def batch_backward(model, info_dict):
|
| 265 |
+
if info_dict["train_engine"] == "deepspeed":
|
| 266 |
+
scaled_loss = model.backward(info_dict["loss_dict"]["loss"])
|
| 267 |
+
else:
|
| 268 |
+
scaled_loss = info_dict["loss_dict"]["loss"] / info_dict["accum_grad"]
|
| 269 |
+
scaled_loss.backward()
|
| 270 |
+
|
| 271 |
+
info_dict["loss_dict"]["loss"] = scaled_loss
|
| 272 |
+
return info_dict
|
| 273 |
+
|
| 274 |
+
|
| 275 |
+
def update_parameter_and_lr(model, optimizer, scheduler, info_dict):
|
| 276 |
+
grad_norm = 0.0
|
| 277 |
+
if info_dict["train_engine"] == "deepspeed":
|
| 278 |
+
info_dict["is_gradient_accumulation_boundary"] = (
|
| 279 |
+
model.is_gradient_accumulation_boundary()
|
| 280 |
+
)
|
| 281 |
+
model.step()
|
| 282 |
+
grad_norm = model.get_global_grad_norm()
|
| 283 |
+
elif (info_dict["batch_idx"] + 1) % info_dict["accum_grad"] == 0:
|
| 284 |
+
grad_norm = clip_grad_norm_(model.parameters(), info_dict["grad_clip"])
|
| 285 |
+
if torch.isfinite(grad_norm):
|
| 286 |
+
optimizer.step()
|
| 287 |
+
optimizer.zero_grad()
|
| 288 |
+
scheduler.step()
|
| 289 |
+
info_dict["lr"] = optimizer.param_groups[0]["lr"]
|
| 290 |
+
info_dict["grad_norm"] = grad_norm
|
| 291 |
+
return info_dict
|
| 292 |
+
|
| 293 |
+
|
| 294 |
+
def log_per_step(writer, info_dict):
|
| 295 |
+
tag = info_dict["tag"]
|
| 296 |
+
epoch = info_dict.get("epoch", 0)
|
| 297 |
+
step = info_dict["step"]
|
| 298 |
+
batch_idx = info_dict["batch_idx"]
|
| 299 |
+
loss_dict = info_dict["loss_dict"]
|
| 300 |
+
rank = int(os.environ.get("RANK", 0))
|
| 301 |
+
|
| 302 |
+
# only rank 0 write to tensorboard to avoid multi-process write
|
| 303 |
+
if writer is not None:
|
| 304 |
+
if (
|
| 305 |
+
info_dict["train_engine"] == "deepspeed"
|
| 306 |
+
and info_dict["is_gradient_accumulation_boundary"] is True
|
| 307 |
+
) or (
|
| 308 |
+
info_dict["train_engine"] == "torch_ddp"
|
| 309 |
+
and (info_dict["batch_idx"] + 1) % info_dict["accum_grad"] == 0
|
| 310 |
+
):
|
| 311 |
+
for k in ["epoch", "lr", "grad_norm"]:
|
| 312 |
+
writer.add_scalar("{}/{}".format(tag, k), info_dict[k], step + 1)
|
| 313 |
+
for k, v in loss_dict.items():
|
| 314 |
+
writer.add_scalar("{}/{}".format(tag, k), v, step + 1)
|
| 315 |
+
|
| 316 |
+
# TRAIN & CV, Shell log (stdout)
|
| 317 |
+
if (info_dict["batch_idx"] + 1) % info_dict["log_interval"] == 0:
|
| 318 |
+
log_str = "{} Batch {}/{} ".format(tag, epoch, batch_idx + 1)
|
| 319 |
+
for name, value in loss_dict.items():
|
| 320 |
+
log_str += "{} {:.6f} ".format(name, value)
|
| 321 |
+
if tag == "TRAIN":
|
| 322 |
+
log_str += "lr {:.8f} grad_norm {:.6f}".format(
|
| 323 |
+
info_dict["lr"], info_dict["grad_norm"]
|
| 324 |
+
)
|
| 325 |
+
log_str += " rank {}".format(rank)
|
| 326 |
+
logging.debug(log_str)
|
| 327 |
+
|
| 328 |
+
|
| 329 |
+
def log_per_save(writer, info_dict):
|
| 330 |
+
tag = info_dict["tag"]
|
| 331 |
+
epoch = info_dict["epoch"]
|
| 332 |
+
step = info_dict["step"]
|
| 333 |
+
loss_dict = info_dict["loss_dict"]
|
| 334 |
+
lr = info_dict["lr"]
|
| 335 |
+
rank = int(os.environ.get("RANK", 0))
|
| 336 |
+
logging.info(
|
| 337 |
+
"Epoch {} Step {} CV info lr {} {} rank {}".format(
|
| 338 |
+
epoch,
|
| 339 |
+
step + 1,
|
| 340 |
+
lr,
|
| 341 |
+
rank,
|
| 342 |
+
" ".join(["{}_{}".format(k, v) for k, v in loss_dict.items()]),
|
| 343 |
+
)
|
| 344 |
+
)
|
| 345 |
+
|
| 346 |
+
if writer is not None:
|
| 347 |
+
for k in ["epoch", "lr"]:
|
| 348 |
+
writer.add_scalar("{}/{}".format(tag, k), info_dict[k], step + 1)
|
| 349 |
+
for k, v in loss_dict.items():
|
| 350 |
+
writer.add_scalar("{}/{}".format(tag, k), v, step + 1)
|
funasr_detach/__init__.py
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Initialize funasr package."""
|
| 2 |
+
|
| 3 |
+
import os
|
| 4 |
+
import pkgutil
|
| 5 |
+
import importlib
|
| 6 |
+
|
| 7 |
+
dirname = os.path.dirname(__file__)
|
| 8 |
+
version_file = os.path.join(dirname, "version.txt")
|
| 9 |
+
with open(version_file, "r") as f:
|
| 10 |
+
__version__ = f.read().strip()
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
import importlib
|
| 14 |
+
import pkgutil
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
def import_submodules(package, recursive=True):
|
| 18 |
+
if isinstance(package, str):
|
| 19 |
+
package = importlib.import_module(package)
|
| 20 |
+
results = {}
|
| 21 |
+
for loader, name, is_pkg in pkgutil.walk_packages(
|
| 22 |
+
package.__path__, package.__name__ + "."
|
| 23 |
+
):
|
| 24 |
+
try:
|
| 25 |
+
results[name] = importlib.import_module(name)
|
| 26 |
+
except Exception as e:
|
| 27 |
+
# 如果想要看到导入错误的具体信息,可以取消注释下面的行
|
| 28 |
+
# print(f"Failed to import {name}: {e}")
|
| 29 |
+
pass
|
| 30 |
+
if recursive and is_pkg:
|
| 31 |
+
results.update(import_submodules(name))
|
| 32 |
+
return results
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
import_submodules(__name__)
|
| 36 |
+
|
| 37 |
+
from funasr_detach.auto.auto_model import AutoModel
|
| 38 |
+
from funasr_detach.auto.auto_frontend import AutoFrontend
|
funasr_detach/auto/__init__.py
ADDED
|
File without changes
|
funasr_detach/auto/auto_frontend.py
ADDED
|
@@ -0,0 +1,90 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import time
|
| 2 |
+
import logging
|
| 3 |
+
from tqdm import tqdm
|
| 4 |
+
|
| 5 |
+
from funasr_detach.register import tables
|
| 6 |
+
from funasr_detach.download.download_from_hub import download_model
|
| 7 |
+
from funasr_detach.utils.load_utils import load_audio_text_image_video, extract_fbank
|
| 8 |
+
from funasr_detach.auto.auto_model import prepare_data_iterator
|
| 9 |
+
from funasr_detach.auto.auto_model import prepare_data_iterator
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
class AutoFrontend:
|
| 13 |
+
def __init__(self, **kwargs):
|
| 14 |
+
assert "model" in kwargs
|
| 15 |
+
if "model_conf" not in kwargs:
|
| 16 |
+
logging.info(
|
| 17 |
+
"download models from model hub: {}".format(
|
| 18 |
+
kwargs.get("model_hub", "ms")
|
| 19 |
+
)
|
| 20 |
+
)
|
| 21 |
+
kwargs = download_model(**kwargs)
|
| 22 |
+
|
| 23 |
+
# build frontend
|
| 24 |
+
frontend = kwargs.get("frontend", None)
|
| 25 |
+
if frontend is not None:
|
| 26 |
+
frontend_class = tables.frontend_classes.get(frontend)
|
| 27 |
+
frontend = frontend_class(**kwargs["frontend_conf"])
|
| 28 |
+
|
| 29 |
+
self.frontend = frontend
|
| 30 |
+
if "frontend" in kwargs:
|
| 31 |
+
del kwargs["frontend"]
|
| 32 |
+
self.kwargs = kwargs
|
| 33 |
+
|
| 34 |
+
def __call__(self, input, input_len=None, kwargs=None, **cfg):
|
| 35 |
+
|
| 36 |
+
kwargs = self.kwargs if kwargs is None else kwargs
|
| 37 |
+
kwargs.update(cfg)
|
| 38 |
+
|
| 39 |
+
key_list, data_list = prepare_data_iterator(input, input_len=input_len)
|
| 40 |
+
batch_size = kwargs.get("batch_size", 1)
|
| 41 |
+
device = kwargs.get("device", "cpu")
|
| 42 |
+
if device == "cpu":
|
| 43 |
+
batch_size = 1
|
| 44 |
+
|
| 45 |
+
meta_data = {}
|
| 46 |
+
|
| 47 |
+
result_list = []
|
| 48 |
+
num_samples = len(data_list)
|
| 49 |
+
pbar = tqdm(colour="blue", total=num_samples + 1, dynamic_ncols=True)
|
| 50 |
+
|
| 51 |
+
time0 = time.perf_counter()
|
| 52 |
+
for beg_idx in range(0, num_samples, batch_size):
|
| 53 |
+
end_idx = min(num_samples, beg_idx + batch_size)
|
| 54 |
+
data_batch = data_list[beg_idx:end_idx]
|
| 55 |
+
key_batch = key_list[beg_idx:end_idx]
|
| 56 |
+
|
| 57 |
+
# extract fbank feats
|
| 58 |
+
time1 = time.perf_counter()
|
| 59 |
+
audio_sample_list = load_audio_text_image_video(
|
| 60 |
+
data_batch, fs=self.frontend.fs, audio_fs=kwargs.get("fs", 16000)
|
| 61 |
+
)
|
| 62 |
+
time2 = time.perf_counter()
|
| 63 |
+
meta_data["load_data"] = f"{time2 - time1:0.3f}"
|
| 64 |
+
speech, speech_lengths = extract_fbank(
|
| 65 |
+
audio_sample_list,
|
| 66 |
+
data_type=kwargs.get("data_type", "sound"),
|
| 67 |
+
frontend=self.frontend,
|
| 68 |
+
**kwargs,
|
| 69 |
+
)
|
| 70 |
+
time3 = time.perf_counter()
|
| 71 |
+
meta_data["extract_feat"] = f"{time3 - time2:0.3f}"
|
| 72 |
+
meta_data["batch_data_time"] = (
|
| 73 |
+
speech_lengths.sum().item()
|
| 74 |
+
* self.frontend.frame_shift
|
| 75 |
+
* self.frontend.lfr_n
|
| 76 |
+
/ 1000
|
| 77 |
+
)
|
| 78 |
+
|
| 79 |
+
speech.to(device=device), speech_lengths.to(device=device)
|
| 80 |
+
batch = {"input": speech, "input_len": speech_lengths, "key": key_batch}
|
| 81 |
+
result_list.append(batch)
|
| 82 |
+
|
| 83 |
+
pbar.update(1)
|
| 84 |
+
description = f"{meta_data}, "
|
| 85 |
+
pbar.set_description(description)
|
| 86 |
+
|
| 87 |
+
time_end = time.perf_counter()
|
| 88 |
+
pbar.set_description(f"time escaped total: {time_end - time0:0.3f}")
|
| 89 |
+
|
| 90 |
+
return result_list
|
funasr_detach/auto/auto_model.py
ADDED
|
@@ -0,0 +1,573 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import json
|
| 2 |
+
import time
|
| 3 |
+
import copy
|
| 4 |
+
import torch
|
| 5 |
+
import random
|
| 6 |
+
import string
|
| 7 |
+
import logging
|
| 8 |
+
import os.path
|
| 9 |
+
import numpy as np
|
| 10 |
+
from tqdm import tqdm
|
| 11 |
+
|
| 12 |
+
from funasr_detach.register import tables
|
| 13 |
+
from funasr_detach.utils.load_utils import load_bytes
|
| 14 |
+
from funasr_detach.download.file import download_from_url
|
| 15 |
+
from funasr_detach.download.download_from_hub import download_model
|
| 16 |
+
from funasr_detach.utils.vad_utils import slice_padding_audio_samples
|
| 17 |
+
from funasr_detach.train_utils.set_all_random_seed import set_all_random_seed
|
| 18 |
+
from funasr_detach.train_utils.load_pretrained_model import load_pretrained_model
|
| 19 |
+
from funasr_detach.utils.load_utils import load_audio_text_image_video
|
| 20 |
+
from funasr_detach.utils.timestamp_tools import timestamp_sentence
|
| 21 |
+
from funasr_detach.models.campplus.utils import sv_chunk, postprocess, distribute_spk
|
| 22 |
+
|
| 23 |
+
try:
|
| 24 |
+
from funasr_detach.models.campplus.cluster_backend import ClusterBackend
|
| 25 |
+
except:
|
| 26 |
+
print("If you want to use the speaker diarization, please `pip install hdbscan`")
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
def prepare_data_iterator(data_in, input_len=None, data_type=None, key=None):
|
| 30 |
+
"""
|
| 31 |
+
|
| 32 |
+
:param input:
|
| 33 |
+
:param input_len:
|
| 34 |
+
:param data_type:
|
| 35 |
+
:param frontend:
|
| 36 |
+
:return:
|
| 37 |
+
"""
|
| 38 |
+
data_list = []
|
| 39 |
+
key_list = []
|
| 40 |
+
filelist = [".scp", ".txt", ".json", ".jsonl"]
|
| 41 |
+
|
| 42 |
+
chars = string.ascii_letters + string.digits
|
| 43 |
+
if isinstance(data_in, str) and data_in.startswith("http"): # url
|
| 44 |
+
data_in = download_from_url(data_in)
|
| 45 |
+
if isinstance(data_in, str) and os.path.exists(
|
| 46 |
+
data_in
|
| 47 |
+
): # wav_path; filelist: wav.scp, file.jsonl;text.txt;
|
| 48 |
+
_, file_extension = os.path.splitext(data_in)
|
| 49 |
+
file_extension = file_extension.lower()
|
| 50 |
+
if file_extension in filelist: # filelist: wav.scp, file.jsonl;text.txt;
|
| 51 |
+
with open(data_in, encoding="utf-8") as fin:
|
| 52 |
+
for line in fin:
|
| 53 |
+
key = "rand_key_" + "".join(random.choice(chars) for _ in range(13))
|
| 54 |
+
if data_in.endswith(
|
| 55 |
+
".jsonl"
|
| 56 |
+
): # file.jsonl: json.dumps({"source": data})
|
| 57 |
+
lines = json.loads(line.strip())
|
| 58 |
+
data = lines["source"]
|
| 59 |
+
key = data["key"] if "key" in data else key
|
| 60 |
+
else: # filelist, wav.scp, text.txt: id \t data or data
|
| 61 |
+
lines = line.strip().split(maxsplit=1)
|
| 62 |
+
data = lines[1] if len(lines) > 1 else lines[0]
|
| 63 |
+
key = lines[0] if len(lines) > 1 else key
|
| 64 |
+
|
| 65 |
+
data_list.append(data)
|
| 66 |
+
key_list.append(key)
|
| 67 |
+
else:
|
| 68 |
+
key = "rand_key_" + "".join(random.choice(chars) for _ in range(13))
|
| 69 |
+
data_list = [data_in]
|
| 70 |
+
key_list = [key]
|
| 71 |
+
elif isinstance(data_in, (list, tuple)):
|
| 72 |
+
if data_type is not None and isinstance(
|
| 73 |
+
data_type, (list, tuple)
|
| 74 |
+
): # mutiple inputs
|
| 75 |
+
data_list_tmp = []
|
| 76 |
+
for data_in_i, data_type_i in zip(data_in, data_type):
|
| 77 |
+
key_list, data_list_i = prepare_data_iterator(
|
| 78 |
+
data_in=data_in_i, data_type=data_type_i
|
| 79 |
+
)
|
| 80 |
+
data_list_tmp.append(data_list_i)
|
| 81 |
+
data_list = []
|
| 82 |
+
for item in zip(*data_list_tmp):
|
| 83 |
+
data_list.append(item)
|
| 84 |
+
else:
|
| 85 |
+
# [audio sample point, fbank, text]
|
| 86 |
+
data_list = data_in
|
| 87 |
+
key_list = [
|
| 88 |
+
"rand_key_" + "".join(random.choice(chars) for _ in range(13))
|
| 89 |
+
for _ in range(len(data_in))
|
| 90 |
+
]
|
| 91 |
+
else: # raw text; audio sample point, fbank; bytes
|
| 92 |
+
if isinstance(data_in, bytes): # audio bytes
|
| 93 |
+
data_in = load_bytes(data_in)
|
| 94 |
+
if key is None:
|
| 95 |
+
key = "rand_key_" + "".join(random.choice(chars) for _ in range(13))
|
| 96 |
+
data_list = [data_in]
|
| 97 |
+
key_list = [key]
|
| 98 |
+
|
| 99 |
+
return key_list, data_list
|
| 100 |
+
|
| 101 |
+
|
| 102 |
+
class AutoModel:
|
| 103 |
+
|
| 104 |
+
def __init__(self, **kwargs):
|
| 105 |
+
if not kwargs.get("disable_log", False):
|
| 106 |
+
tables.print()
|
| 107 |
+
|
| 108 |
+
model, kwargs = self.build_model(**kwargs)
|
| 109 |
+
|
| 110 |
+
# if vad_model is not None, build vad model else None
|
| 111 |
+
vad_model = kwargs.get("vad_model", None)
|
| 112 |
+
vad_kwargs = kwargs.get("vad_model_revision", None)
|
| 113 |
+
if vad_model is not None:
|
| 114 |
+
logging.info("Building VAD model.")
|
| 115 |
+
vad_kwargs = {
|
| 116 |
+
"model": vad_model,
|
| 117 |
+
"model_revision": vad_kwargs,
|
| 118 |
+
"device": kwargs["device"],
|
| 119 |
+
}
|
| 120 |
+
vad_model, vad_kwargs = self.build_model(**vad_kwargs)
|
| 121 |
+
|
| 122 |
+
# if punc_model is not None, build punc model else None
|
| 123 |
+
punc_model = kwargs.get("punc_model", None)
|
| 124 |
+
punc_kwargs = kwargs.get("punc_model_revision", None)
|
| 125 |
+
if punc_model is not None:
|
| 126 |
+
logging.info("Building punc model.")
|
| 127 |
+
punc_kwargs = {
|
| 128 |
+
"model": punc_model,
|
| 129 |
+
"model_revision": punc_kwargs,
|
| 130 |
+
"device": kwargs["device"],
|
| 131 |
+
}
|
| 132 |
+
punc_model, punc_kwargs = self.build_model(**punc_kwargs)
|
| 133 |
+
|
| 134 |
+
# if spk_model is not None, build spk model else None
|
| 135 |
+
spk_model = kwargs.get("spk_model", None)
|
| 136 |
+
spk_kwargs = kwargs.get("spk_model_revision", None)
|
| 137 |
+
if spk_model is not None:
|
| 138 |
+
logging.info("Building SPK model.")
|
| 139 |
+
spk_kwargs = {
|
| 140 |
+
"model": spk_model,
|
| 141 |
+
"model_revision": spk_kwargs,
|
| 142 |
+
"device": kwargs["device"],
|
| 143 |
+
}
|
| 144 |
+
spk_model, spk_kwargs = self.build_model(**spk_kwargs)
|
| 145 |
+
self.cb_model = ClusterBackend().to(kwargs["device"])
|
| 146 |
+
spk_mode = kwargs.get("spk_mode", "punc_segment")
|
| 147 |
+
if spk_mode not in ["default", "vad_segment", "punc_segment"]:
|
| 148 |
+
logging.error(
|
| 149 |
+
"spk_mode should be one of default, vad_segment and punc_segment."
|
| 150 |
+
)
|
| 151 |
+
self.spk_mode = spk_mode
|
| 152 |
+
|
| 153 |
+
self.kwargs = kwargs
|
| 154 |
+
self.model = model
|
| 155 |
+
self.vad_model = vad_model
|
| 156 |
+
self.vad_kwargs = vad_kwargs
|
| 157 |
+
self.punc_model = punc_model
|
| 158 |
+
self.punc_kwargs = punc_kwargs
|
| 159 |
+
self.spk_model = spk_model
|
| 160 |
+
self.spk_kwargs = spk_kwargs
|
| 161 |
+
self.model_path = kwargs.get("model_path")
|
| 162 |
+
|
| 163 |
+
def build_model(self, **kwargs):
|
| 164 |
+
assert "model" in kwargs
|
| 165 |
+
if "model_conf" not in kwargs:
|
| 166 |
+
logging.info(
|
| 167 |
+
"download models from model hub: {}".format(
|
| 168 |
+
kwargs.get("model_hub", "ms")
|
| 169 |
+
)
|
| 170 |
+
)
|
| 171 |
+
kwargs = download_model(**kwargs)
|
| 172 |
+
|
| 173 |
+
set_all_random_seed(kwargs.get("seed", 0))
|
| 174 |
+
|
| 175 |
+
device = kwargs.get("device", "cuda")
|
| 176 |
+
if not torch.cuda.is_available() or kwargs.get("ngpu", 1) == 0:
|
| 177 |
+
device = "cpu"
|
| 178 |
+
kwargs["batch_size"] = 1
|
| 179 |
+
kwargs["device"] = device
|
| 180 |
+
|
| 181 |
+
if kwargs.get("ncpu", None):
|
| 182 |
+
torch.set_num_threads(kwargs.get("ncpu"))
|
| 183 |
+
|
| 184 |
+
# build tokenizer
|
| 185 |
+
tokenizer = kwargs.get("tokenizer", None)
|
| 186 |
+
if tokenizer is not None:
|
| 187 |
+
tokenizer_class = tables.tokenizer_classes.get(tokenizer)
|
| 188 |
+
tokenizer = tokenizer_class(**kwargs["tokenizer_conf"])
|
| 189 |
+
kwargs["tokenizer"] = tokenizer
|
| 190 |
+
kwargs["token_list"] = tokenizer.token_list
|
| 191 |
+
vocab_size = len(tokenizer.token_list)
|
| 192 |
+
else:
|
| 193 |
+
vocab_size = -1
|
| 194 |
+
|
| 195 |
+
# build frontend
|
| 196 |
+
frontend = kwargs.get("frontend", None)
|
| 197 |
+
if frontend is not None:
|
| 198 |
+
frontend_class = tables.frontend_classes.get(frontend)
|
| 199 |
+
frontend = frontend_class(**kwargs["frontend_conf"])
|
| 200 |
+
kwargs["frontend"] = frontend
|
| 201 |
+
kwargs["input_size"] = frontend.output_size()
|
| 202 |
+
|
| 203 |
+
# build model
|
| 204 |
+
model_class = tables.model_classes.get(kwargs["model"])
|
| 205 |
+
model = model_class(**kwargs, **kwargs["model_conf"], vocab_size=vocab_size)
|
| 206 |
+
|
| 207 |
+
model.to(device)
|
| 208 |
+
|
| 209 |
+
# init_param
|
| 210 |
+
init_param = kwargs.get("init_param", None)
|
| 211 |
+
if init_param is not None:
|
| 212 |
+
logging.info(f"Loading pretrained params from {init_param}")
|
| 213 |
+
load_pretrained_model(
|
| 214 |
+
model=model,
|
| 215 |
+
path=init_param,
|
| 216 |
+
ignore_init_mismatch=kwargs.get("ignore_init_mismatch", False),
|
| 217 |
+
oss_bucket=kwargs.get("oss_bucket", None),
|
| 218 |
+
scope_map=kwargs.get("scope_map", None),
|
| 219 |
+
excludes=kwargs.get("excludes", None),
|
| 220 |
+
)
|
| 221 |
+
|
| 222 |
+
return model, kwargs
|
| 223 |
+
|
| 224 |
+
def __call__(self, *args, **cfg):
|
| 225 |
+
kwargs = self.kwargs
|
| 226 |
+
kwargs.update(cfg)
|
| 227 |
+
res = self.model(*args, kwargs)
|
| 228 |
+
return res
|
| 229 |
+
|
| 230 |
+
def generate(self, input, input_len=None, **cfg):
|
| 231 |
+
if self.vad_model is None:
|
| 232 |
+
return self.inference(input, input_len=input_len, **cfg)
|
| 233 |
+
|
| 234 |
+
else:
|
| 235 |
+
return self.inference_with_vad(input, input_len=input_len, **cfg)
|
| 236 |
+
|
| 237 |
+
def inference(
|
| 238 |
+
self, input, input_len=None, model=None, kwargs=None, key=None, **cfg
|
| 239 |
+
):
|
| 240 |
+
kwargs = self.kwargs if kwargs is None else kwargs
|
| 241 |
+
kwargs.update(cfg)
|
| 242 |
+
model = self.model if model is None else model
|
| 243 |
+
model = model.cuda()
|
| 244 |
+
model.eval()
|
| 245 |
+
|
| 246 |
+
batch_size = kwargs.get("batch_size", 1)
|
| 247 |
+
# if kwargs.get("device", "cpu") == "cpu":
|
| 248 |
+
# batch_size = 1
|
| 249 |
+
|
| 250 |
+
key_list, data_list = prepare_data_iterator(
|
| 251 |
+
input, input_len=input_len, data_type=kwargs.get("data_type", None), key=key
|
| 252 |
+
)
|
| 253 |
+
|
| 254 |
+
speed_stats = {}
|
| 255 |
+
asr_result_list = []
|
| 256 |
+
num_samples = len(data_list)
|
| 257 |
+
disable_pbar = kwargs.get("disable_pbar", False)
|
| 258 |
+
pbar = (
|
| 259 |
+
tqdm(colour="blue", total=num_samples, dynamic_ncols=True)
|
| 260 |
+
if not disable_pbar
|
| 261 |
+
else None
|
| 262 |
+
)
|
| 263 |
+
time_speech_total = 0.0
|
| 264 |
+
time_escape_total = 0.0
|
| 265 |
+
for beg_idx in range(0, num_samples, batch_size):
|
| 266 |
+
end_idx = min(num_samples, beg_idx + batch_size)
|
| 267 |
+
data_batch = data_list[beg_idx:end_idx]
|
| 268 |
+
key_batch = key_list[beg_idx:end_idx]
|
| 269 |
+
batch = {"data_in": data_batch, "key": key_batch}
|
| 270 |
+
if (end_idx - beg_idx) == 1 and kwargs.get(
|
| 271 |
+
"data_type", None
|
| 272 |
+
) == "fbank": # fbank
|
| 273 |
+
batch["data_in"] = data_batch[0]
|
| 274 |
+
batch["data_lengths"] = input_len
|
| 275 |
+
|
| 276 |
+
time1 = time.perf_counter()
|
| 277 |
+
with torch.no_grad():
|
| 278 |
+
results, meta_data = model.inference(**batch, **kwargs)
|
| 279 |
+
time2 = time.perf_counter()
|
| 280 |
+
|
| 281 |
+
asr_result_list.extend(results)
|
| 282 |
+
|
| 283 |
+
# batch_data_time = time_per_frame_s * data_batch_i["speech_lengths"].sum().item()
|
| 284 |
+
batch_data_time = meta_data.get("batch_data_time", -1)
|
| 285 |
+
time_escape = time2 - time1
|
| 286 |
+
speed_stats["load_data"] = meta_data.get("load_data", 0.0)
|
| 287 |
+
speed_stats["extract_feat"] = meta_data.get("extract_feat", 0.0)
|
| 288 |
+
speed_stats["forward"] = f"{time_escape:0.3f}"
|
| 289 |
+
speed_stats["batch_size"] = f"{len(results)}"
|
| 290 |
+
speed_stats["time_cost"] = f"{(time_escape)}"
|
| 291 |
+
speed_stats["rtf"] = f"{(time_escape) / batch_data_time:0.3f}"
|
| 292 |
+
description = f"{speed_stats}, "
|
| 293 |
+
if pbar:
|
| 294 |
+
pbar.update(1)
|
| 295 |
+
pbar.set_description(description)
|
| 296 |
+
time_speech_total += batch_data_time
|
| 297 |
+
time_escape_total += time_escape
|
| 298 |
+
|
| 299 |
+
if pbar:
|
| 300 |
+
# pbar.update(1)
|
| 301 |
+
pbar.set_description(f"rtf_avg: {time_escape_total/time_speech_total:0.3f}")
|
| 302 |
+
torch.cuda.empty_cache()
|
| 303 |
+
return asr_result_list
|
| 304 |
+
|
| 305 |
+
def inference_with_vad(self, input, input_len=None, **cfg):
|
| 306 |
+
|
| 307 |
+
# step.1: compute the vad model
|
| 308 |
+
self.vad_kwargs.update(cfg)
|
| 309 |
+
beg_vad = time.time()
|
| 310 |
+
res = self.inference(
|
| 311 |
+
input,
|
| 312 |
+
input_len=input_len,
|
| 313 |
+
model=self.vad_model,
|
| 314 |
+
kwargs=self.vad_kwargs,
|
| 315 |
+
**cfg,
|
| 316 |
+
)
|
| 317 |
+
end_vad = time.time()
|
| 318 |
+
print(f"time cost vad: {end_vad - beg_vad:0.3f}")
|
| 319 |
+
|
| 320 |
+
# step.2 compute asr model
|
| 321 |
+
model = self.model
|
| 322 |
+
kwargs = self.kwargs
|
| 323 |
+
kwargs.update(cfg)
|
| 324 |
+
batch_size = int(kwargs.get("batch_size_s", 300)) * 1000
|
| 325 |
+
batch_size_threshold_ms = int(kwargs.get("batch_size_threshold_s", 60)) * 1000
|
| 326 |
+
kwargs["batch_size"] = batch_size
|
| 327 |
+
|
| 328 |
+
key_list, data_list = prepare_data_iterator(
|
| 329 |
+
input, input_len=input_len, data_type=kwargs.get("data_type", None)
|
| 330 |
+
)
|
| 331 |
+
results_ret_list = []
|
| 332 |
+
time_speech_total_all_samples = 1e-6
|
| 333 |
+
|
| 334 |
+
beg_total = time.time()
|
| 335 |
+
pbar_total = tqdm(colour="red", total=len(res), dynamic_ncols=True)
|
| 336 |
+
for i in range(len(res)):
|
| 337 |
+
key = res[i]["key"]
|
| 338 |
+
vadsegments = res[i]["value"]
|
| 339 |
+
input_i = data_list[i]
|
| 340 |
+
speech = load_audio_text_image_video(
|
| 341 |
+
input_i, fs=kwargs["frontend"].fs, audio_fs=kwargs.get("fs", 16000)
|
| 342 |
+
)
|
| 343 |
+
speech_lengths = len(speech)
|
| 344 |
+
n = len(vadsegments)
|
| 345 |
+
data_with_index = [(vadsegments[i], i) for i in range(n)]
|
| 346 |
+
sorted_data = sorted(data_with_index, key=lambda x: x[0][1] - x[0][0])
|
| 347 |
+
results_sorted = []
|
| 348 |
+
|
| 349 |
+
if not len(sorted_data):
|
| 350 |
+
logging.info("decoding, utt: {}, empty speech".format(key))
|
| 351 |
+
continue
|
| 352 |
+
|
| 353 |
+
if len(sorted_data) > 0 and len(sorted_data[0]) > 0:
|
| 354 |
+
batch_size = max(
|
| 355 |
+
batch_size, sorted_data[0][0][1] - sorted_data[0][0][0]
|
| 356 |
+
)
|
| 357 |
+
|
| 358 |
+
batch_size_ms_cum = 0
|
| 359 |
+
beg_idx = 0
|
| 360 |
+
beg_asr_total = time.time()
|
| 361 |
+
time_speech_total_per_sample = speech_lengths / 16000
|
| 362 |
+
time_speech_total_all_samples += time_speech_total_per_sample
|
| 363 |
+
|
| 364 |
+
all_segments = []
|
| 365 |
+
for j, _ in enumerate(range(0, n)):
|
| 366 |
+
# pbar_sample.update(1)
|
| 367 |
+
batch_size_ms_cum += sorted_data[j][0][1] - sorted_data[j][0][0]
|
| 368 |
+
if (
|
| 369 |
+
j < n - 1
|
| 370 |
+
and (
|
| 371 |
+
batch_size_ms_cum
|
| 372 |
+
+ sorted_data[j + 1][0][1]
|
| 373 |
+
- sorted_data[j + 1][0][0]
|
| 374 |
+
)
|
| 375 |
+
< batch_size
|
| 376 |
+
and (sorted_data[j + 1][0][1] - sorted_data[j + 1][0][0])
|
| 377 |
+
< batch_size_threshold_ms
|
| 378 |
+
):
|
| 379 |
+
continue
|
| 380 |
+
batch_size_ms_cum = 0
|
| 381 |
+
end_idx = j + 1
|
| 382 |
+
speech_j, speech_lengths_j = slice_padding_audio_samples(
|
| 383 |
+
speech, speech_lengths, sorted_data[beg_idx:end_idx]
|
| 384 |
+
)
|
| 385 |
+
results = self.inference(
|
| 386 |
+
speech_j,
|
| 387 |
+
input_len=None,
|
| 388 |
+
model=model,
|
| 389 |
+
kwargs=kwargs,
|
| 390 |
+
disable_pbar=True,
|
| 391 |
+
**cfg,
|
| 392 |
+
)
|
| 393 |
+
if self.spk_model is not None:
|
| 394 |
+
# compose vad segments: [[start_time_sec, end_time_sec, speech], [...]]
|
| 395 |
+
for _b in range(len(speech_j)):
|
| 396 |
+
vad_segments = [
|
| 397 |
+
[
|
| 398 |
+
sorted_data[beg_idx:end_idx][_b][0][0] / 1000.0,
|
| 399 |
+
sorted_data[beg_idx:end_idx][_b][0][1] / 1000.0,
|
| 400 |
+
np.array(speech_j[_b]),
|
| 401 |
+
]
|
| 402 |
+
]
|
| 403 |
+
segments = sv_chunk(vad_segments)
|
| 404 |
+
all_segments.extend(segments)
|
| 405 |
+
speech_b = [i[2] for i in segments]
|
| 406 |
+
spk_res = self.inference(
|
| 407 |
+
speech_b,
|
| 408 |
+
input_len=None,
|
| 409 |
+
model=self.spk_model,
|
| 410 |
+
kwargs=kwargs,
|
| 411 |
+
disable_pbar=True,
|
| 412 |
+
**cfg,
|
| 413 |
+
)
|
| 414 |
+
results[_b]["spk_embedding"] = spk_res[0]["spk_embedding"]
|
| 415 |
+
beg_idx = end_idx
|
| 416 |
+
if len(results) < 1:
|
| 417 |
+
continue
|
| 418 |
+
results_sorted.extend(results)
|
| 419 |
+
|
| 420 |
+
restored_data = [0] * n
|
| 421 |
+
for j in range(n):
|
| 422 |
+
index = sorted_data[j][1]
|
| 423 |
+
restored_data[index] = results_sorted[j]
|
| 424 |
+
result = {}
|
| 425 |
+
|
| 426 |
+
# results combine for texts, timestamps, speaker embeddings and others
|
| 427 |
+
# TODO: rewrite for clean code
|
| 428 |
+
for j in range(n):
|
| 429 |
+
for k, v in restored_data[j].items():
|
| 430 |
+
if k.startswith("timestamp"):
|
| 431 |
+
if k not in result:
|
| 432 |
+
result[k] = []
|
| 433 |
+
for t in restored_data[j][k]:
|
| 434 |
+
t[0] += vadsegments[j][0]
|
| 435 |
+
t[1] += vadsegments[j][0]
|
| 436 |
+
result[k].extend(restored_data[j][k])
|
| 437 |
+
elif k == "spk_embedding":
|
| 438 |
+
if k not in result:
|
| 439 |
+
result[k] = restored_data[j][k]
|
| 440 |
+
else:
|
| 441 |
+
result[k] = torch.cat(
|
| 442 |
+
[result[k], restored_data[j][k]], dim=0
|
| 443 |
+
)
|
| 444 |
+
elif "text" in k:
|
| 445 |
+
if k not in result:
|
| 446 |
+
result[k] = restored_data[j][k]
|
| 447 |
+
else:
|
| 448 |
+
result[k] += " " + restored_data[j][k]
|
| 449 |
+
else:
|
| 450 |
+
if k not in result:
|
| 451 |
+
result[k] = restored_data[j][k]
|
| 452 |
+
else:
|
| 453 |
+
result[k] += restored_data[j][k]
|
| 454 |
+
|
| 455 |
+
return_raw_text = kwargs.get("return_raw_text", False)
|
| 456 |
+
# step.3 compute punc model
|
| 457 |
+
if self.punc_model is not None:
|
| 458 |
+
self.punc_kwargs.update(cfg)
|
| 459 |
+
punc_res = self.inference(
|
| 460 |
+
result["text"],
|
| 461 |
+
model=self.punc_model,
|
| 462 |
+
kwargs=self.punc_kwargs,
|
| 463 |
+
disable_pbar=True,
|
| 464 |
+
**cfg,
|
| 465 |
+
)
|
| 466 |
+
raw_text = copy.copy(result["text"])
|
| 467 |
+
if return_raw_text:
|
| 468 |
+
result["raw_text"] = raw_text
|
| 469 |
+
result["text"] = punc_res[0]["text"]
|
| 470 |
+
else:
|
| 471 |
+
raw_text = None
|
| 472 |
+
|
| 473 |
+
# speaker embedding cluster after resorted
|
| 474 |
+
if self.spk_model is not None and kwargs.get("return_spk_res", True):
|
| 475 |
+
if raw_text is None:
|
| 476 |
+
logging.error("Missing punc_model, which is required by spk_model.")
|
| 477 |
+
all_segments = sorted(all_segments, key=lambda x: x[0])
|
| 478 |
+
spk_embedding = result["spk_embedding"]
|
| 479 |
+
labels = self.cb_model(
|
| 480 |
+
spk_embedding.cpu(), oracle_num=kwargs.get("preset_spk_num", None)
|
| 481 |
+
)
|
| 482 |
+
# del result['spk_embedding']
|
| 483 |
+
sv_output = postprocess(all_segments, None, labels, spk_embedding.cpu())
|
| 484 |
+
if self.spk_mode == "vad_segment": # recover sentence_list
|
| 485 |
+
sentence_list = []
|
| 486 |
+
for res, vadsegment in zip(restored_data, vadsegments):
|
| 487 |
+
if "timestamp" not in res:
|
| 488 |
+
logging.error(
|
| 489 |
+
"Only 'iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch' \
|
| 490 |
+
and 'iic/speech_seaco_paraformer_large_asr_nat-zh-cn-16k-common-vocab8404-pytorch'\
|
| 491 |
+
can predict timestamp, and speaker diarization relies on timestamps."
|
| 492 |
+
)
|
| 493 |
+
sentence_list.append(
|
| 494 |
+
{
|
| 495 |
+
"start": vadsegment[0],
|
| 496 |
+
"end": vadsegment[1],
|
| 497 |
+
"sentence": res["text"],
|
| 498 |
+
"timestamp": res["timestamp"],
|
| 499 |
+
}
|
| 500 |
+
)
|
| 501 |
+
elif self.spk_mode == "punc_segment":
|
| 502 |
+
if "timestamp" not in result:
|
| 503 |
+
logging.error(
|
| 504 |
+
"Only 'iic/speech_paraformer-large-vad-punc_asr_nat-zh-cn-16k-common-vocab8404-pytorch' \
|
| 505 |
+
and 'iic/speech_seaco_paraformer_large_asr_nat-zh-cn-16k-common-vocab8404-pytorch'\
|
| 506 |
+
can predict timestamp, and speaker diarization relies on timestamps."
|
| 507 |
+
)
|
| 508 |
+
sentence_list = timestamp_sentence(
|
| 509 |
+
punc_res[0]["punc_array"],
|
| 510 |
+
result["timestamp"],
|
| 511 |
+
raw_text,
|
| 512 |
+
return_raw_text=return_raw_text,
|
| 513 |
+
)
|
| 514 |
+
distribute_spk(sentence_list, sv_output)
|
| 515 |
+
result["sentence_info"] = sentence_list
|
| 516 |
+
elif kwargs.get("sentence_timestamp", False):
|
| 517 |
+
sentence_list = timestamp_sentence(
|
| 518 |
+
punc_res[0]["punc_array"],
|
| 519 |
+
result["timestamp"],
|
| 520 |
+
raw_text,
|
| 521 |
+
return_raw_text=return_raw_text,
|
| 522 |
+
)
|
| 523 |
+
result["sentence_info"] = sentence_list
|
| 524 |
+
if "spk_embedding" in result:
|
| 525 |
+
del result["spk_embedding"]
|
| 526 |
+
|
| 527 |
+
result["key"] = key
|
| 528 |
+
results_ret_list.append(result)
|
| 529 |
+
end_asr_total = time.time()
|
| 530 |
+
time_escape_total_per_sample = end_asr_total - beg_asr_total
|
| 531 |
+
pbar_total.update(1)
|
| 532 |
+
pbar_total.set_description(
|
| 533 |
+
f"rtf_avg: {time_escape_total_per_sample / time_speech_total_per_sample:0.3f}, "
|
| 534 |
+
f"time_speech: {time_speech_total_per_sample: 0.3f}, "
|
| 535 |
+
f"time_escape: {time_escape_total_per_sample:0.3f}"
|
| 536 |
+
)
|
| 537 |
+
|
| 538 |
+
return results_ret_list
|
| 539 |
+
|
| 540 |
+
def infer_encoder(
|
| 541 |
+
self, input, input_len=None, model=None, kwargs=None, key=None, **cfg
|
| 542 |
+
):
|
| 543 |
+
kwargs = self.kwargs if kwargs is None else kwargs
|
| 544 |
+
kwargs.update(cfg)
|
| 545 |
+
model = self.model if model is None else model
|
| 546 |
+
model = model.cuda()
|
| 547 |
+
model.eval()
|
| 548 |
+
|
| 549 |
+
batch_size = kwargs.get("batch_size", 1)
|
| 550 |
+
|
| 551 |
+
key_list, data_list = prepare_data_iterator(
|
| 552 |
+
input, input_len=input_len, data_type=kwargs.get("data_type", None), key=key
|
| 553 |
+
)
|
| 554 |
+
|
| 555 |
+
asr_result_list = []
|
| 556 |
+
num_samples = len(data_list)
|
| 557 |
+
for beg_idx in range(0, num_samples, batch_size):
|
| 558 |
+
end_idx = min(num_samples, beg_idx + batch_size)
|
| 559 |
+
data_batch = data_list[beg_idx:end_idx]
|
| 560 |
+
key_batch = key_list[beg_idx:end_idx]
|
| 561 |
+
batch = {"data_in": data_batch, "key": key_batch}
|
| 562 |
+
if (end_idx - beg_idx) == 1 and kwargs.get(
|
| 563 |
+
"data_type", None
|
| 564 |
+
) == "fbank": # fbank
|
| 565 |
+
batch["data_in"] = data_batch[0]
|
| 566 |
+
batch["data_lengths"] = input_len
|
| 567 |
+
|
| 568 |
+
with torch.no_grad():
|
| 569 |
+
results, meta_data, cache = model.infer_encoder(**batch, **kwargs)
|
| 570 |
+
asr_result_list.extend(results)
|
| 571 |
+
|
| 572 |
+
torch.cuda.empty_cache()
|
| 573 |
+
return asr_result_list, cache
|
funasr_detach/auto/auto_tokenizer.py
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
class AutoTokenizer:
|
| 2 |
+
"""
|
| 3 |
+
Undo
|
| 4 |
+
"""
|
| 5 |
+
|
| 6 |
+
def __init__(self):
|
| 7 |
+
pass
|