Spaces:
Paused
Paused
Update app_g.py
Browse files
app_g.py
CHANGED
@@ -10,6 +10,42 @@ import numpy as np
|
|
10 |
from diffsynth import ModelManager, PusaMultiFramesPipeline, PusaV2VPipeline, WanVideoPusaPipeline, save_video
|
11 |
import tempfile
|
12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
class PusaVideoDemo:
|
14 |
def __init__(self):
|
15 |
self.device = "cuda" if torch.cuda.is_available() else "cpu"
|
@@ -38,6 +74,8 @@ class PusaVideoDemo:
|
|
38 |
torch_dtype=torch.bfloat16,
|
39 |
)
|
40 |
print("Models loaded successfully!")
|
|
|
|
|
41 |
|
42 |
def load_lora_and_get_pipe(self, pipe_type, lora_path, lora_alpha):
|
43 |
"""Load LoRA and return appropriate pipeline"""
|
@@ -244,7 +282,7 @@ class PusaVideoDemo:
|
|
244 |
|
245 |
except Exception as e:
|
246 |
return None, f"Error: {str(e)}"
|
247 |
-
|
248 |
def generate_t2v_video(self, prompt, lora_alpha, num_inference_steps,
|
249 |
negative_prompt, progress=gr.Progress()):
|
250 |
"""Generate video from text prompt"""
|
@@ -1191,6 +1229,7 @@ def create_demo():
|
|
1191 |
return demo
|
1192 |
|
1193 |
if __name__ == "__main__":
|
|
|
1194 |
demo = create_demo()
|
1195 |
demo.launch(
|
1196 |
share=False,
|
|
|
10 |
from diffsynth import ModelManager, PusaMultiFramesPipeline, PusaV2VPipeline, WanVideoPusaPipeline, save_video
|
11 |
import tempfile
|
12 |
|
13 |
+
|
14 |
+
|
15 |
+
# Constants
|
16 |
+
import os
|
17 |
+
from huggingface_hub import snapshot_download
|
18 |
+
|
19 |
+
# Constants
|
20 |
+
MODEL_ZOO_DIR = "./model_zoo"
|
21 |
+
PUSA_DIR = os.path.join(MODEL_ZOO_DIR, "PusaV1")
|
22 |
+
WAN_SUBFOLDER = "Wan2.1-T2V-14B"
|
23 |
+
WAN_MODEL_PATH = os.path.join(PUSA_DIR, WAN_SUBFOLDER)
|
24 |
+
LORA_PATH = os.path.join(PUSA_DIR, "pusa_v1.pt")
|
25 |
+
|
26 |
+
# Ensure model and weights are downloaded
|
27 |
+
def ensure_model_downloaded():
|
28 |
+
if not os.path.exists(PUSA_DIR):
|
29 |
+
print("Downloading RaphaelLiu/PusaV1 to ./model_zoo/PusaV1 ...")
|
30 |
+
snapshot_download(
|
31 |
+
repo_id="RaphaelLiu/PusaV1",
|
32 |
+
local_dir=PUSA_DIR,
|
33 |
+
repo_type="model",
|
34 |
+
local_dir_use_symlinks=False,
|
35 |
+
)
|
36 |
+
print("✅ PusaV1 downloaded.")
|
37 |
+
|
38 |
+
if not os.path.exists(WAN_MODEL_PATH):
|
39 |
+
print("Downloading Wan-AI/Wan2.1-T2V-14B to ./model_zoo/PusaV1/Wan2.1-T2V-14B ...")
|
40 |
+
snapshot_download(
|
41 |
+
repo_id="Wan-AI/Wan2.1-T2V-14B",
|
42 |
+
local_dir=WAN_MODEL_PATH, # Changed from WAN_DIR to WAN_MODEL_PATH
|
43 |
+
repo_type="model",
|
44 |
+
local_dir_use_symlinks=False,
|
45 |
+
)
|
46 |
+
print("✅ Wan2.1-T2V-14B downloaded.")
|
47 |
+
|
48 |
+
|
49 |
class PusaVideoDemo:
|
50 |
def __init__(self):
|
51 |
self.device = "cuda" if torch.cuda.is_available() else "cpu"
|
|
|
74 |
torch_dtype=torch.bfloat16,
|
75 |
)
|
76 |
print("Models loaded successfully!")
|
77 |
+
|
78 |
+
|
79 |
|
80 |
def load_lora_and_get_pipe(self, pipe_type, lora_path, lora_alpha):
|
81 |
"""Load LoRA and return appropriate pipeline"""
|
|
|
282 |
|
283 |
except Exception as e:
|
284 |
return None, f"Error: {str(e)}"
|
285 |
+
|
286 |
def generate_t2v_video(self, prompt, lora_alpha, num_inference_steps,
|
287 |
negative_prompt, progress=gr.Progress()):
|
288 |
"""Generate video from text prompt"""
|
|
|
1229 |
return demo
|
1230 |
|
1231 |
if __name__ == "__main__":
|
1232 |
+
ensure_model_downloaded()
|
1233 |
demo = create_demo()
|
1234 |
demo.launch(
|
1235 |
share=False,
|