四川住房建设厅网站首页,搭建网页游戏,网站 展示板,做汽车拆解视频网站时长比较短的音频#xff1a;https://huggingface.co/datasets/PolyAI/minds14/viewer/en-US
时长比较长的音频#xff1a;https://huggingface.co/datasets/librispeech_asr?row8
此次测试过程暂时只使用比较短的音频 使用fast_whisper测试
下载安装#xff0c;参考官方…
时长比较短的音频https://huggingface.co/datasets/PolyAI/minds14/viewer/en-US
时长比较长的音频https://huggingface.co/datasets/librispeech_asr?row8
此次测试过程暂时只使用比较短的音频 使用fast_whisper测试
下载安装参考官方网站即可 报错提示 Could not load library libcudnn_ops_infer.so.8. Error: libcudnn_ops_infer.so.8: cannot open shared object file: No such file or directory Please make sure libcudnn_ops_infer.so.8 is in your library path! 解决办法
找到有libcudnn_ops_infer.so.8 的路径在我的电脑中改文件所在的路径为 在终端导入 export LD_LIBRARY_PATH/opt/audio/venv/lib/python3.10/site-packages/nvidia/cudnn/lib:$LD_LIBRARY_PATH
test_fast_whisper.py import subprocess
import os
import time
import unittest
import openpyxl
from pydub import AudioSegment
from datasets import load_datasetfrom faster_whisper import WhisperModelclass TestFastWhisper(unittest.TestCase):def setUp(self):passdef test_fastwhisper(self):# 替换为您的脚本路径# 设置HTTP代理os.environ[http_proxy] http://10.10.10.178:7890os.environ[HTTP_PROXY] http://10.10.10.178:7890# 不知道此处为什么不能生效必须要在终端中手动导入os.environ[LD_LIBRARY_PATH] /opt/audio/venv/lib/python3.10/site-packages/nvidia/cudnn/lib:$LD_LIBRARY_PATH# 设置HTTPS代理os.environ[https_proxy] http://10.10.10.178:7890os.environ[HTTPS_PROXY] http://10.10.10.178:7890print(load whisper)# 使用fast_whisper model_size large-v2# Run on GPU with FP16fast_whisper_model WhisperModel(model_size, devicecuda, compute_typefloat16)minds_14 load_dataset(PolyAI/minds14, en-US, splittrain) # for en-USworkbook openpyxl.Workbook()# 创建一个工作表worksheet workbook.active# 设置表头worksheet[A1] Audio Pathworksheet[B1] Audio Duration (seconds)worksheet[C1] Audio Size (MB)worksheet[D1] Correct Textworksheet[E1] Transcribed Textworksheet[F1] Cost Time (seconds)for index, each in enumerate(minds_14, start2):audioPath each[path]print(audioPath)# audioArray each[audio]audioDuration len(AudioSegment.from_file(audioPath))/1000audioSize os.path.getsize(audioPath)/ (1024 * 1024)CorrectText each[transcription]tran_start_time time.time()segments, info fast_whisper_model.transcribe(audioPath, beam_size5)segments list(segments) # The transcription will actually run here.print(Detected language %s with probability %f % (info.language, info.language_probability))text for segment in segments:text segment.textcost_time time.time() - tran_start_timeprint(Audio Path:, audioPath)print(Audio Duration (seconds):, audioDuration)print(Audio Size (MB):, audioSize)print(Correct Text:, CorrectText)print(Transcription Time (seconds):, cost_time)print(Transcribed Text:, text)worksheet[fA{index}] audioPathworksheet[fB{index}] audioDurationworksheet[fC{index}] audioSizeworksheet[fD{index}] CorrectTextworksheet[fE{index}] textworksheet[fF{index}] cost_time# breakworkbook.save(fast_whisper_output_data.xlsx)print(数据已保存到 fast_whisper_output_data.xlsx 文件)if __name__ __main__:unittest.main() 使用whisper测试
下载安装参考官方网站即可代码与上面代码类似 测试结果可视化
不太熟悉用numbers凑合着看一下就行 很明显fast_whisper速度要更快一些
