Update app.py

app.py

@@ -8,32 +8,12 @@ from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
 from huggingface_hub import hf_hub_download
 from datasets import load_dataset
 from whisper import load_model, transcribe
+from faster_whisper import WhisperModel
 
 
 @spaces.GPU
-def transcribe_audio(zip_file, progress = gr.Progress(track_tqdm= True)):
-    device = "cuda:0" if torch.cuda.is_available() else "cpu"
-    torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
-    
-    model_id = "distil-whisper/distil-large-v3"
-    
-    model = AutoModelForSpeechSeq2Seq.from_pretrained(
-        model_id, torch_dtype=torch_dtype, low_cpu_mem_usage=True, use_safetensors=True
-    )
-    model.to(device)
-    
-    processor = AutoProcessor.from_pretrained(model_id)
-    
-    pipe = pipeline(
-        "automatic-speech-recognition",
-        model=model,
-        tokenizer=processor.tokenizer,
-        feature_extractor=processor.feature_extractor,
-        max_new_tokens=128,
-        torch_dtype=torch_dtype,
-        device=device,
-        generate_kwargs={"task": "transcribe"}
-    )
+def transcribe_audio(zip_file, progress = gr.Progress(track_tqdm= True)):    
+    model = WhisperModel("large-v3")
       
     
     # Create a temporary directory to extract the ZIP file
@@ -48,7 +28,9 @@ def transcribe_audio(zip_file, progress = gr.Progress(track_tqdm= True)):
         # Transcribe each audio file
         transcriptions = {}
         for audio_file in audio_files:
-            transcription = pipe(audio_file)["text"]
+                segments, info = model.transcribe("audio.mp3")
+                for segment in segments:
+                    return "[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text)
             transcriptions[os.path.basename(audio_file)] = transcription
         
     return transcriptions