Spaces:
Sleeping
Sleeping
File size: 1,162 Bytes
dcf6735 a976d7a 16d11ec dcf6735 16d11ec dcf6735 8ef7225 dcf6735 16d11ec |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 |
import torch
from transformers import Speech2TextProcessor, Speech2TextForConditionalGeneration
from audio_recorder_streamlit import audio_recorder
import numpy as np
import streamlit as st
def transcribe_audio(audio_bytes):
model = Speech2TextForConditionalGeneration.from_pretrained("facebook/s2t-small-mustc-en-fr-st")
processor = Speech2TextProcessor.from_pretrained("facebook/s2t-small-mustc-en-fr-st")
# Convert audio bytes to tensors
input_features = torch.tensor(audio_bytes).unsqueeze(0) # Assuming audio_bytes is numpy array
# Generate transcription
generated_ids = model.generate(input_features)
translation = processor.batch_decode(generated_ids, skip_special_tokens=True)
return translation
st.title("Audio to Text Transcription..")
audio_bytes = audio_recorder(pause_threshold=3.0, sample_rate=16_000)
if audio_bytes:
st.audio(audio_bytes, format="audio/wav")
transcription = transcribe_audio(audio_bytes)
if transcription:
st.write("Transcription:")
st.write(transcription)
else:
st.write("Error: Failed to transcribe audio.")
else:
st.write("No audio recorded.")
|