Spaces:
Sleeping
Sleeping
import gradio as gr | |
from transformers import Wav2Vec2ForCTC, AutoProcessor | |
import torch | |
import librosa | |
import json | |
with open('ISO_codes.json', 'r') as file: | |
iso_codes = json.load(file) | |
model_id = "TifinLab/mms-1b-berber" | |
processor = AutoProcessor.from_pretrained(model_id) | |
model = Wav2Vec2ForCTC.from_pretrained(model_id) | |
def transcribe(audio_file_mic=None, audio_file_upload=None): | |
if audio_file_mic: | |
audio_file = audio_file_mic | |
elif audio_file_upload: | |
audio_file = audio_file_upload | |
else: | |
return "Please upload an audio file or record one" | |
# Make sure audio is 16kHz | |
speech, sample_rate = librosa.load(audio_file) | |
if sample_rate != 16000: | |
speech = librosa.resample(speech, orig_sr=sample_rate, target_sr=16000) | |
processor.tokenizer.set_target_lang("ber") | |
model.load_adapter("ber") | |
inputs = processor(speech, sampling_rate=16_000, return_tensors="pt") | |
with torch.no_grad(): | |
outputs = model(**inputs).logits | |
ids = torch.argmax(outputs, dim=-1)[0] | |
transcription = processor.decode(ids) | |
return transcription | |
examples = [["kab_1.mp3", None, "Amazigh (kab)"], | |
["kab_2.mp3", None, "Amazigh (kab)"]] | |
description = '' | |
iface = gr.Interface(fn=transcribe, | |
inputs=[ | |
gr.Audio(type="filepath", label="Enregistrez ou téléchargez votre réponse audio ici") | |
], | |
outputs=gr.Textbox(label="Transcription"), | |
examples=examples, | |
description=description | |
) | |
iface.launch() |