Spaces:
Runtime error
Runtime error
| import gradio as gr | |
| import os | |
| from transformers import WhisperProcessor, WhisperForConditionalGeneration | |
| import numpy as np | |
| import librosa | |
| # Initialize Whisper model | |
| processor = WhisperProcessor.from_pretrained("openai/whisper-base") | |
| model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-base") | |
| # Set light green theme | |
| theme = gr.themes.Base( | |
| primary_hue="emerald", | |
| secondary_hue="emerald", | |
| neutral_hue="gray", | |
| ) | |
| def validate_file(file_path): | |
| # Check if file exists | |
| if not file_path or not os.path.exists(file_path): | |
| return False, "No file uploaded or file not found." | |
| # Check file size (25 MB limit) | |
| file_size_mb = os.path.getsize(file_path) / (1024 * 1024) | |
| if file_size_mb > 25: | |
| return False, f"File size is {file_size_mb:.2f} MB. Please upload a file smaller than 25 MB." | |
| # Check file extension | |
| file_extension = os.path.splitext(file_path)[1].lower() | |
| if file_extension not in ['.mp3', '.wav']: | |
| return False, "Only .mp3 and .wav formats are supported." | |
| return True, "File is valid." | |
| def transcribe_audio(audio_file): | |
| # Check if audio_file is None | |
| if audio_file is None: | |
| return "Please upload an audio file." | |
| # Validate the file first | |
| is_valid, message = validate_file(audio_file) | |
| if not is_valid: | |
| return message | |
| try: | |
| # Load audio file | |
| speech_array, sampling_rate = librosa.load(audio_file, sr=16000) | |
| # Process the audio file | |
| input_features = processor(speech_array, sampling_rate=16000, return_tensors="pt").input_features | |
| # Generate token ids | |
| predicted_ids = model.generate(input_features) | |
| # Decode token ids to text | |
| transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0] | |
| return transcription | |
| except Exception as e: | |
| return f"An error occurred during transcription: {str(e)}" | |
| # Create Gradio interface | |
| with gr.Blocks(theme=theme) as demo: | |
| gr.Markdown("# Audio Transcription with Whisper") | |
| gr.Markdown("Upload an audio file (.mp3 or .wav) of maximum 25MB to get the transcription.") | |
| with gr.Row(): | |
| with gr.Column(): | |
| # Fixed: Use sources parameter instead of type | |
| audio_input = gr.Audio(sources=["upload"], label="Upload Audio File") | |
| submit_btn = gr.Button("Transcribe", variant="primary") | |
| with gr.Column(): | |
| output = gr.Textbox(label="Transcription Result", lines=10) | |
| submit_btn.click(fn=transcribe_audio, inputs=audio_input, outputs=output) | |
| gr.Markdown("### Limitations") | |
| gr.Markdown("- Maximum file size: 25 MB") | |
| gr.Markdown("- Supported formats: .mp3 and .wav") | |
| gr.Markdown("- Uses the Whisper base model which works best with clear audio") | |
| # Launch the app | |
| if __name__ == "__main__": | |
| demo.launch() |