blog

app.py

@@ -68,7 +68,7 @@ def transcribe_audio(audio_file):
     inputs = ffmpeg_read(inputs, asr_pipeline.feature_extractor.sampling_rate)
     inputs = {"array": inputs, "sampling_rate": asr_pipeline.feature_extractor.sampling_rate}
 
-    transcription = asr_pipeline(inputs, batch_size=8, generate_kwargs={"task": "transcribe"}, return_timestamps=True)["text"]
+    transcription = asr_pipeline(inputs, batch_size=8, generate_kwargs={"task": "translate"}, return_timestamps=False)["text"]
     return transcription
 
 @spaces.GPU
@@ -91,7 +91,7 @@ def analyze_video(prompt, video, progress=gr.Progress()):
     
     progress(0.5, desc="Transcribing audio")
     transcription = transcribe_audio(audio_path)
-    progress(f"Transcription: {transcription}")
+    print(f"Transcription: {transcription}")
     
     # Clean up temporary audio file
     os.remove(audio_path)