使用whisper转文本,这里使用的是large-v3版本
pip install git+https://github.com/openai/whisper.git
import whisper
import os
from tqdm import tqdmmodel = whisper.load_model("large-v3")path = "rag_data"
for fi in tqdm(os.listdir(path)):file = os.path.join(path, fi)result = model.transcribe(file)text = result["text"]with open(os.path.join("rad_all_texts", file.split('.')[0] + ".txt"), 'a') as f:f.write(text + '\n')
参考
Introducing Whisper
OpenAI Whisper 新一代语音技术(更新至v3)
fine-tune-whisper
Whisper-Finetune