app_dcase.py requirements.txt: go back to simple demo with CLAP twice, place dcase baseline in dedicated branch until installation is solved

app_dcase.py

@@ -1,40 +1,26 @@
 
-# import gradio as gr
-# from msclap import CLAP
-
-# clap_model = CLAP(version = 'clapcap', use_cuda=False)
-
-# def clap_inference(mic=None, file=None):
-
-#     if mic is not None:
-#         audio = mic
-#     elif file is not None:
-#         audio = file
-#     else:
-#         return "You must either provide a mic recording or a file"
+import gradio as gr
+from msclap import CLAP
 
-#     # Generate captions for the recording
-#     captions = clap_model.generate_caption([audio], 
-#                                            resample=True, 
-#                                            beam_size=5, 
-#                                            entry_length=67, 
-#                                            temperature=0.01)
+clap_model = CLAP(version = 'clapcap', use_cuda=False)
 
-#     return captions[0]
+def clap_inference(mic=None, file=None):
 
-from transformers import pipeline
-import gradio as gr
-import torchaudio
-import torch
-from dcase24t6.nn.hub import baseline_pipeline
-import os
-import gdown
+    if mic is not None:
+        audio = mic
+    elif file is not None:
+        audio = file
+    else:
+        return "You must either provide a mic recording or a file"
 
-def download_dcase_model_checkpoint():
+    # Generate captions for the recording
+    captions = clap_model.generate_caption([audio], 
+                                           resample=True, 
+                                           beam_size=5, 
+                                           entry_length=67, 
+                                           temperature=0.01)
 
-    url = "https://drive.google.com/uc?id=1JABWIBlHuLAhYPX5ktbyLniH-YpeRyeT"
-    output = 'epoch_232-step_001864-mode_min-val_loss_3.3752.ckpt'
-    gdown.download(url, output, quiet=False)
+    return captions[0]
 
 def create_app():
 
@@ -45,7 +31,7 @@ def create_app():
             """
         )
         gr.Interface(
-            fn=dcase_inference,
+            fn=clap_inference,
             inputs=[
                 gr.Audio(sources="microphone", type="filepath"),
                 gr.Audio(sources="upload", type="filepath"),
@@ -55,8 +41,6 @@ def create_app():
 
     return demo
 
-download_dcase_model_checkpoint()
-
 def main():
     
     app = create_app()

requirements.txt

@@ -1,8 +1,2 @@
 gradio==5.13.1
 msclap
-# transformers
-# torch==2.2.1
-# torchoutil[extras]~=0.3.0
-# ffmpeg
-# ffmpeg-python
-git+https://github.com/Labbeti/dcase2024-task6-baseline