Update app.py

app.py

@@ -1,16 +1,27 @@
 import gradio as gr
+from transformers import WhisperProcessor, WhisperForConditionalGeneration
 
-# Load the model without launching the interface
-loaded_model = gr.Interface.load("models/openai/whisper-large-v2", allow_launch=False)
+# Load model and processor
+processor = WhisperProcessor.from_pretrained("openai/whisper-large-v2")
+model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-large-v2")
 
-def transcribe_audio(audio_file):
-    # Use the loaded model to transcribe the audio
-    return loaded_model(audio_file)
+def transcribe_audio(audio_path: str) -> str:
+    with open(audio_path, "rb") as f:
+        audio_data = f.read()
+    
+    # Get audio features
+    input_features = processor(audio_data, return_tensors="pt").input_features 
 
-audio_input = gr.inputs.Audio(type="filepath")
-text_output = gr.outputs.Textbox()
+    # Transcribe without forcing any context tokens so that the model tries to automatically detect the language
+    model.config.forced_decoder_ids = None
+    predicted_ids = model.generate(input_features)
+    transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
+    
+    return transcription[0]
+
+audio_input = gr.inputs.Audio(type="file", label="Upload an audio file")
+text_output = gr.outputs.Textbox(label="Transcription")
 
-# Setup the custom Gradio interface with your configurations
 iface = gr.Interface(
     fn=transcribe_audio,
     inputs=audio_input,