Update requirement

app.py

@@ -2,19 +2,24 @@ import os
 import gradio as gr
 import torchaudio
 import torch
+import numpy as np
 from pipeline.morph_pipeline_successed_ver1 import AudioLDM2MorphPipeline 
 
+# Initialize AudioLDM2 Pipeline
 pipeline = AudioLDM2MorphPipeline.from_pretrained("cvssp/audioldm2-large", torch_dtype=torch.float32)
 pipeline.to("cuda")
 
+# Audio morphing function
 def morph_audio(audio_file1, audio_file2, prompt1, prompt2, negative_prompt1="Low quality", negative_prompt2="Low quality"):
     save_lora_dir = "output"
     os.makedirs(save_lora_dir, exist_ok=True)
     
+    # Load audio and compute duration
     waveform, sample_rate = torchaudio.load(audio_file1)
     duration = waveform.shape[1] / sample_rate
     duration = int(duration)
     
+    # Perform morphing using the pipeline
     _ = pipeline(
         audio_file=audio_file1,
         audio_file2=audio_file2,
@@ -27,7 +32,7 @@ def morph_audio(audio_file1, audio_file2, prompt1, prompt2, negative_prompt1="Lo
         negative_prompt_2=negative_prompt2,
         save_lora_dir=save_lora_dir,
         use_adain=True,
-        use_reschedule=True,
+        use_reschedule=False,
         num_inference_steps=50,
         lamd=0.6,
         output_path=save_lora_dir,
@@ -41,32 +46,33 @@ def morph_audio(audio_file1, audio_file2, prompt1, prompt2, negative_prompt1="Lo
         guidance_scale=7.5,
     )
     
+    # Collect the output file paths
     output_paths = [os.path.join(save_lora_dir, file) for file in os.listdir(save_lora_dir) if file.endswith(".wav")]
     return output_paths
 
+# Gradio interface function
 def interface(audio1, audio2, prompt1, prompt2):
     output_paths = morph_audio(audio1, audio2, prompt1, prompt2)
-    return output_paths 
+    return output_paths
 
-# Gradio UI
-with gr.Blocks() as demo:
-    gr.Markdown("### Audio Morphing Demo with AudioLDM2")
-    
-    with gr.Row():
-        audio_file1 = gr.Audio(label="Upload Audio File 1", type="filepath")
-        audio_file2 = gr.Audio(label="Upload Audio File 2", type="filepath")
-    
-    with gr.Row():
-        prompt1 = gr.Textbox(label="Prompt for Audio File 1")
-        prompt2 = gr.Textbox(label="Prompt for Audio File 2")
-    
-    output_audios = gr.Audio(label="Generated Morphing Audios", type="filepath", interactive=False)
-    morph_button = gr.Button("Generate Morphing Audio")
-    
-    morph_button.click(
-        interface, 
-        inputs=[audio_file1, audio_file2, prompt1, prompt2], 
-        outputs=[output_audios]
-    )
+# Gradio Interface
+demo = gr.Interface(
+    fn=interface,
+    inputs=[
+        gr.Audio(label="Upload Audio File 1", type="filepath"),
+        gr.Audio(label="Upload Audio File 2", type="filepath"),
+        # gr.Slider(4, 6, step=1, label="Octave 1"),
+        gr.Textbox(label="Prompt for Audio File 1"),
+        gr.Textbox(label="Prompt for Audio File 2")
+    ],
+    outputs=[
+        gr.Audio(label="Generated Tone 1"),
+        gr.Audio(label="Generated Tone 2"),
+        gr.Audio(label="Generated Tone 3"),
+        gr.Audio(label="Generated Tone 4"),
+        gr.Audio(label="Generated Tone 5"),
+    ],
+)
 
-demo.launch()
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -76,3 +76,4 @@ uvicorn==0.32.1
 wavaugment==0.2
 websockets==12.0
 zstandard==0.23.0
+timm