nl8590687 / ASRT_SpeechRecognition

A Deep-Learning-Based Chinese Speech Recognition System 基于深度学习的中文语音识别系统
https://asrt.ailemon.net
GNU General Public License v3.0
7.85k stars 1.9k forks source link

字错误率100%,loss下降很慢 #203

Open wxhiff opened 4 years ago

wxhiff commented 4 years ago

大家好,我在Google Cloab上运行python train_mspeech.py,一直得到语音单字错误率和dev单字错误率为100%,而且loss到了210左右下降很慢,请问正常吗?

*[测试结果]语音识别dev集语音单字错误率:100.0% [message epodh. Have train datas11000+ Epoch 1/1 500/500[ ========================]-145s291ms/step-loss:209.9455 测试进度:0/4 *[测试结果]语音识别 train集语音单字错误率:100.0% 测试进度:0/4 *[测试结果]语音识别dev集语音单字错误率:100.0% [message] epoch 0. Have train datas 11500+ Epoch 1/1 500/500[=====================]-144s288ms/step-loss:210.5319 测试进度:0/4 *[测试结果]语音识别train集语音单字错误率:100.0% 测试进度:0/4 *[测试结果]语音识别dev集语音单字错误率:100.0% [message] epoch 0. Have train datas 12000+ Epoch 1/1 500/500[======================]-144s288ms/step-loss:209.1676 测试进度:0/4 *[测试结果]语音识别 train集语音单错误率:100.0% 测试进度:0/4 *[测试结果]语音识别dev集语音单字错误率:100.0% [message] epoch 0. Have train datas 12500+ Epoch 1/1 500/500[=========================]-143s285ms/step-loss:209.7521 测试进度:0/4 *[测试结果]语音识别train集语音单字错误率:100.0% 测试进度:0/4 *[测试结果]语音识别dev集语音单字错误率:100.0% [message] epoch 0. Have train datas 13000+ Epoch 1/1 227/500 8-1.2065

xc005 commented 3 years ago

我也遇到了相同的问题,同问正常吗?

drzqb commented 3 years ago

同问啊,怎么回事?

nl8590687 commented 3 years ago

没在Google Cloab上运行过,请先验证运行环境是否有问题

nl8590687 commented 3 years ago

没在Google Cloab上运行过,请先验证运行环境是否有问题