Spaces:
Runtime error
Runtime error
| # # #uvicorn app:app --host 0.0.0.0 --port 8000 --reload | |
| # # # from fastapi import FastAPI | |
| # # # from transformers import WhisperProcessor, WhisperForConditionalGeneration | |
| # # # import librosa | |
| # # # import uvicorn | |
| # # # app = FastAPI() | |
| # # # processor = WhisperProcessor.from_pretrained("openai/whisper-small") | |
| # # # model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small") | |
| # # # model.config.forced_decoder_ids = None | |
| # # # audio_file_path = "output.mp3" | |
| # # # audio_data, _ = librosa.load(audio_file_path, sr=16000) | |
| # # # @app.get("/") | |
| # # # def transcribe_audio(): | |
| # # # input_features = processor(audio_data.tolist(), return_tensors="pt").input_features | |
| # # # predicted_ids = model.generate(input_features) | |
| # # # transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) | |
| # # # return {"transcription": transcription[0]} | |
| # # # if __name__ == "__main__": | |
| # # # import uvicorn | |
| # # # uvicorn.run(app, host="0.0.0.0", port=8000) | |
| # # # if __name__=='__main__': | |
| # # # uvicorn.run('main:app', reload=True) | |
| # # #uvicorn app:app --host 0.0.0.0 --port 8000 --reload | |
| # # #curl -X GET "http://localhost:8000/?text=I%20like%20Apples" | |
| # # #http://localhost:8000/?text=I%20like%20Apples | |
| # # # from fastapi import FastAPI | |
| # # # from transformers import WhisperProcessor, WhisperForConditionalGeneration | |
| # # # import librosa | |
| # # # import uvicorn | |
| # # # app = FastAPI() | |
| # # # # Load model and processor | |
| # # # processor = WhisperProcessor.from_pretrained("openai/whisper-small") | |
| # # # model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small") | |
| # # # model.config.forced_decoder_ids = None | |
| # # # # Path to your audio file | |
| # # # audio_file_path = "/home/pranjal/Downloads/output.mp3" | |
| # # # # Read the audio file | |
| # # # audio_data, _ = librosa.load(audio_file_path, sr=16000) | |
| # # # @app.get("/") | |
| # # # def transcribe_audio(): | |
| # # # # Process the audio data using the Whisper processor | |
| # # # input_features = processor(audio_data.tolist(), return_tensors="pt").input_features | |
| # # # # Generate transcription | |
| # # # predicted_ids = model.generate(input_features) | |
| # # # transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) | |
| # # # return {"transcription": transcription[0]} | |
| # # # if __name__ == "__main__": | |
| # # # import uvicorn | |
| # # # uvicorn.run(app, host="0.0.0.0", port=8000) | |
| # # # if __name__=='__app__': | |
| # # # uvicorn.run('main:app', reload=True) | |
| # # from fastapi import FastAPI, UploadFile, File | |
| # # from transformers import WhisperProcessor, WhisperForConditionalGeneration | |
| # # import librosa | |
| # # from fastapi.responses import HTMLResponse | |
| # # import uvicorn | |
| # # import io | |
| # # app = FastAPI() | |
| # # # Load model and processor | |
| # # processor = WhisperProcessor.from_pretrained("openai/whisper-small") | |
| # # model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small") | |
| # # model.config.forced_decoder_ids = None | |
| # # @app.get("/") | |
| # # def read_root(): | |
| # # html_form = """ | |
| # # <html> | |
| # # <body> | |
| # # <h2>ASR Transcription</h2> | |
| # # <form action="/transcribe" method="post" enctype="multipart/form-data"> | |
| # # <label for="audio_file">Upload an audio file (MP3 or WAV):</label> | |
| # # <input type="file" id="audio_file" name="audio_file" accept=".mp3, .wav" required><br><br> | |
| # # <input type="submit" value="Transcribe"> | |
| # # </form> | |
| # # </body> | |
| # # </html> | |
| # # """ | |
| # # return HTMLResponse(content=html_form, status_code=200) | |
| # # @app.post("/transcribe") | |
| # # async def transcribe_audio(audio_file: UploadFile): | |
| # # try: | |
| # # # Read the uploaded audio file | |
| # # audio_data = await audio_file.read() | |
| # # # Process the audio data using the Whisper processor | |
| # # audio_data, _ = librosa.load(io.BytesIO(audio_data), sr=16000) | |
| # # input_features = processor(audio_data.tolist(), return_tensors="pt").input_features | |
| # # # Generate transcription | |
| # # predicted_ids = model.generate(input_features) | |
| # # transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) | |
| # # return {"transcription": transcription[0]} | |
| # # except Exception as e: | |
| # # return {"error": str(e)} | |
| # # if __name__ == "__app__": | |
| # # uvicorn.run(app, host="0.0.0.0", port=8000, reload=True) | |
| # #uvicorn app:app --host 0.0.0.0 --port 8000 --reload | |
| # from fastapi import FastAPI, UploadFile, File | |
| # from transformers import WhisperProcessor, WhisperForConditionalGeneration | |
| # import librosa | |
| # from fastapi.responses import HTMLResponse | |
| # import uvicorn | |
| # import io | |
| # app = FastAPI() | |
| # # # Load model and processor | |
| # # processor = WhisperProcessor.from_pretrained("openai/whisper-medium") | |
| # # model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-medium") | |
| # # model.config.forced_decoder_ids = None | |
| # import whisper | |
| # model = whisper.load_model("small") | |
| # @app.get("/") | |
| # def read_root(): | |
| # html_form = """ | |
| # <html> | |
| # <body> | |
| # <h2>ASR Transcription</h2> | |
| # <form action="/transcribe" method="post" enctype="multipart/form-data"> | |
| # <label for="audio_file">Upload an audio file (MP3 or WAV):</label> | |
| # <input type="file" id="audio_file" name="audio_file" accept=".mp3, .wav" required><br><br> | |
| # <input type="submit" value="Transcribe"> | |
| # </form> | |
| # </body> | |
| # </html> | |
| # """ | |
| # return HTMLResponse(content=html_form, status_code=200) | |
| # @app.post("/transcribe") | |
| # async def transcribe_audio(audio_file: UploadFile): | |
| # try: | |
| # # Read the uploaded audio file | |
| # audio_data = await audio_file.read() | |
| # # Process the audio data using the Whisper processor | |
| # # audio_data, _ = librosa.load(io.BytesIO(audio_data), sr=16000) | |
| # # input_features = processor(audio_data.tolist(), return_tensors="pt").input_features | |
| # # # Generate transcription | |
| # # predicted_ids = model.generate(input_features) | |
| # # transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) | |
| # result = model.transcribe("/home/pranjal/Downloads/rt.mp3") | |
| # return {"transcription": result['text']} | |
| # except Exception as e: | |
| # return {"error": str(e)} | |
| # # if __name__ == "__app__": | |
| # # uvicorn.run(app, host="0.0.0.0", port=8000, reload=True) | |
| #uvicorn app:app --host 0.0.0.0 --port 8000 --reload | |
| from fastapi import FastAPI, UploadFile, File | |
| from transformers import WhisperProcessor, WhisperForConditionalGeneration | |
| from fastapi.responses import HTMLResponse | |
| import librosa | |
| import io | |
| import re | |
| html_tag_remover = re.compile(r'<[^>]+>') | |
| def remove_tags(text): | |
| return html_tag_remover.sub('', text) | |
| app = FastAPI() | |
| processor = WhisperProcessor.from_pretrained("openai/whisper-small") | |
| model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small") | |
| model.config.forced_decoder_ids = None | |
| chunk_duration = 30 | |
| overlap_duration = 5 | |
| def read_root(): | |
| html_form = """ | |
| <html> | |
| <body> | |
| <h2>ASR Transcription</h2> | |
| <form action="/transcribe" method="post" enctype="multipart/form-data"> | |
| <label for="audio_file">Upload an audio file (MP3 or WAV):</label> | |
| <input type="file" id="audio_file" name="audio_file" accept=".mp3, .wav" required><br><br> | |
| <input type="submit" value="Transcribe"> | |
| </form> | |
| </body> | |
| </html> | |
| """ | |
| return HTMLResponse(content=html_form, status_code=200) | |
| async def transcribe_audio(audio_file: UploadFile): | |
| audio_data = await audio_file.read() | |
| audio_data, _ = librosa.load(io.BytesIO(audio_data), sr=16000) | |
| transcription = [] | |
| start = 0 | |
| while start < len(audio_data): | |
| end = start + chunk_duration * 16000 | |
| audio_chunk = audio_data[start:end] | |
| input_features = processor(audio_chunk.tolist(), return_tensors="pt").input_features | |
| predicted_ids = model.generate(input_features, max_length=1000) | |
| chunk_transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True) | |
| transcription.extend(chunk_transcription) | |
| start = end - overlap_duration * 16000 | |
| final_transcription = " ".join(transcription) | |
| final_transcription = remove_tags(final_transcription) | |
| return {"transcription": final_transcription} | |