Futuresony's picture
Rename app.py to app.py1
cf1f728 verified
import gradio as gr
import soundfile as sf
import torch
from transformers import Wav2Vec2ForCTC, AutoProcessor
# Assuming 'transcribe' was defined in a previous cell.
# If not, define it here or import it from the correct module.
# Create a placeholder for ASR_LANGUAGES if it's not defined elsewhere.
ASR_LANGUAGES = {"eng": "English", "swh": "Swahili"} # Replace with your actual languages
# ✅ Define or Re-define the `transcribe` function within this cell
MODEL_ID = "facebook/mms-1b-all" # Make sure this is the same model ID used for training
processor = AutoProcessor.from_pretrained(MODEL_ID)
model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)
def transcribe(audio_path, language):
"""Transcribes an audio file using the fine-tuned model."""
# Set the target language based on user selection
if language:
target_lang = language.split(" ")[0] # Extract language code
processor.tokenizer.set_target_lang(target_lang)
if target_lang != "eng": # Load adapter if not English
model.load_adapter(target_lang)
audio, samplerate = sf.read(audio_path)
inputs = processor(audio, sampling_rate=samplerate, return_tensors="pt")
with torch.no_grad():
outputs = model(**inputs).logits
ids = torch.argmax(outputs, dim=-1)[0]
return processor.decode(ids)
mms_transcribe = gr.Interface(
fn=transcribe,
inputs=[
gr.Audio(),
gr.Dropdown(
[f"{k} ({v})" for k, v in ASR_LANGUAGES.items()],
label="Language",
value="eng English",
),
],
outputs="text",
title="Speech-to-Text Transcription",
description="Transcribe audio input into text.",
allow_flagging="never",
)
with gr.Blocks() as demo:
mms_transcribe.render()
if __name__ == "__main__":
demo.queue()
demo.launch()