Reduced GPU wait time for testing, removed preemptive Whisper load

app.py

@@ -24,12 +24,12 @@ device = "cuda" if cuda_available else "cpu"
 logger.info(f"Using device: {device}")
 
 # Load Whisper model
-print("Loading Whisper model...")
-try:
-    load_models()  # Load Whisper model
-except Exception as e:
-    logger.error(f"Error loading Whisper model: {str(e)}")
-    raise
+# print("Loading Whisper model...")
+# try:
+#     load_models()  # Load Whisper model
+# except Exception as e:
+#     logger.error(f"Error loading Whisper model: {str(e)}")
+#     raise
 
 print("Whisper model loaded successfully.")
 
@@ -57,11 +57,11 @@ def process_with_fallback(func, *args, **kwargs):
         else:
             raise
 
-@spaces.GPU(duration=600)
+@spaces.GPU(duration=60)
 def transcribe_audio(audio_file, translate, model_size, use_diarization):
     return process_with_fallback(process_audio, audio_file, translate=translate, model_size=model_size, use_diarization=use_diarization)
 
-@spaces.GPU(duration=600)
+@spaces.GPU(duration=60)
 def summarize_text(text):
     summarizer = load_summarization_model()
     try:
@@ -72,7 +72,7 @@ def summarize_text(text):
         summary = "Error occurred during summarization. Please try again."
     return summary
 
-@spaces.GPU(duration=600)
+@spaces.GPU(duration=60)
 def process_and_summarize(audio_file, translate, model_size, use_diarization, do_summarize):
     transcription, full_text = transcribe_audio(audio_file, translate, model_size, use_diarization)
     summary = summarize_text(full_text) if do_summarize else ""
@@ -81,12 +81,12 @@ def process_and_summarize(audio_file, translate, model_size, use_diarization, do
 # Main interface
 with gr.Blocks() as iface:
     gr.Markdown("# WhisperX Audio Transcription, Translation, and Summarization (with ZeroGPU support)")
-    
+
     audio_input = gr.Audio(type="filepath")
     translate_checkbox = gr.Checkbox(label="Enable Translation")
     summarize_checkbox = gr.Checkbox(label="Enable Summarization", interactive=False)
-    model_dropdown = gr.Dropdown(choices=["tiny", "base", "small", "medium", "large", "large-v2", "large-v3"], label="Whisper Model Size", value="small")
     diarization_checkbox = gr.Checkbox(label="Enable Speaker Diarization")
+    model_dropdown = gr.Dropdown(choices=["tiny", "base", "small", "medium", "large", "large-v2", "large-v3"], label="Whisper Model Size", value="small")
     process_button = gr.Button("Process Audio")
     transcription_output = gr.Textbox(label="Transcription/Translation")
     summary_output = gr.Textbox(label="Summary")