Update app.py

app.py

@@ -1,12 +1,13 @@
+from flask import Flask, request, jsonify
 import requests
 import torch
 import librosa
 from transformers import WhisperProcessor, WhisperForConditionalGeneration
-from flask import Flask, request, jsonify
 
 app = Flask(__name__)
 
-model_id = "openai/whisper-large-v3"
+# Temporarily using smaller model for faster testing
+model_id = "openai/whisper-base"
 processor = WhisperProcessor.from_pretrained(model_id)
 model = WhisperForConditionalGeneration.from_pretrained(model_id)
 
@@ -33,27 +34,20 @@ def transcribe_audio(audio_url):
         inputs = processor(chunk, sampling_rate=16000, return_tensors="pt", padding=True)
         input_features = inputs.input_features.to(device)
 
-        with torch.no_grad():
-            predicted_ids = model.generate(
-                input_features,
-                forced_decoder_ids=forced_decoder_ids
-            )
-
+        predicted_ids = model.generate(input_features)
         transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
         partial_text += transcription + "\n"
 
-    return partial_text.strip()
+    return partial_text
 
 @app.route('/transcribe', methods=['POST'])
 def transcribe_endpoint():
     data = request.get_json()
-    audio_url = data.get('audio_url')
-    if not audio_url:
-        return jsonify({"error": "Missing 'audio_url' in request"}), 400
+    audio_url = data['audio_url']
 
     transcription = transcribe_audio(audio_url)
 
-    return jsonify({"transcription": transcription})
+    return {"transcription": transcription}
 
 if __name__ == '__main__':
     app.run(host="0.0.0.0", port=8080)