Update app.py

app.py

@@ -40,10 +40,18 @@ def transcribe(audio_path):
     # predicted_ids = torch.argmax(logits, dim=-1)
     # transcription = processor.batch_decode(predicted_ids)
 
-    transcription = model(audio_path,batch_size=1000, generate_kwargs={"task": "transcribe"}, return_timestamps=True)["text"]
+    # transcription = model(audio_path,batch_size=1000, generate_kwargs={"task": "transcribe"}, return_timestamps=True)["text"]
 
     # result = pipe(sample)
     # 返回转录结果
+    # return transcription
+
+
+    #------
+    audio_input = processor(audio_path, return_tensors="pt", sampling_rate=16000).input_values
+    with torch.no_grad():
+        predicted_ids = model.generate(audio_input)
+    transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
     return transcription