Skip to content

Instantly share code, notes, and snippets.

@GoooIce
Created March 1, 2023 05:32
Show Gist options
  • Select an option

  • Save GoooIce/b22c83122ecd310d02725b38481b9a53 to your computer and use it in GitHub Desktop.

Select an option

Save GoooIce/b22c83122ecd310d02725b38481b9a53 to your computer and use it in GitHub Desktop.
import torch
import gradio as gr
from transformers import (
AutomaticSpeechRecognitionPipeline,
WhisperForConditionalGeneration,
WhisperTokenizer,
WhisperProcessor,
)
from peft import PeftModel, PeftConfig
peft_model_id = "GoooIce/openai-whisper-large-v2-LORA-colab"
language = "Chinese"
task = "transcribe"
peft_config = PeftConfig.from_pretrained(peft_model_id)
model = WhisperForConditionalGeneration.from_pretrained(
peft_config.base_model_name_or_path, load_in_8bit=True, device_map="auto"
)
model = PeftModel.from_pretrained(model, peft_model_id)
tokenizer = WhisperTokenizer.from_pretrained(peft_config.base_model_name_or_path, language=language, task=task)
processor = WhisperProcessor.from_pretrained(peft_config.base_model_name_or_path, language=language, task=task)
feature_extractor = processor.feature_extractor
forced_decoder_ids = processor.get_decoder_prompt_ids(language=language, task=task)
pipe = AutomaticSpeechRecognitionPipeline(model=model, tokenizer=tokenizer, feature_extractor=feature_extractor)
def transcribe(audio):
with torch.cuda.amp.autocast():
text = pipe(audio, generate_kwargs={"forced_decoder_ids": forced_decoder_ids}, max_new_tokens=255)["text"]
return text
iface = gr.Interface(
fn=transcribe,
inputs=gr.Audio(source="microphone", type="filepath"),
outputs="text",
title="PEFT LoRA + INT8 Whisper Large V2 Marathi",
description="Realtime demo for Marathi speech recognition using `PEFT-LoRA+INT8` fine-tuned Whisper Large V2 model.",
)
iface.launch(share=True)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment