deepspeech2_tensorflow

所属分类:模式识别(视觉/语音等)
开发工具:Python
文件大小:40KB
下载次数:0
上传日期:2019-06-25 11:51:42
上 传 者sh-1993
说明:  deepspeech2_tensorflow,基于baidu 2016年开源语音识别项目修改后端为tensorflow,rnn层为 双向GRU,ctc损失计算调用tensorflow内置模块实现中文语音识别,在清华开源数据上的测试准确率0.97.
(deepspeech2_ Tensorflow, based on Baidu s 2016 open-source speech recognition project, modified the backend to Tensorflow, the RNN layer to bidirectional GRU, and the ctc loss calculation calls the Tensorflow built-in module to achieve Chinese speech recognition. The test accuracy on Tsinghua open-source data is 0.97)

文件列表:
char_map.py (256, 2019-06-25)
config_util.py (532, 2019-06-25)
cp_model.sh (259, 2019-06-25)
data_generator.py (13763, 2019-06-25)
dict_4562.txt (39951, 2019-06-25)
model.py (5046, 2019-06-25)
plot_acc.py (3791, 2019-06-25)
run.sh (1808, 2019-06-25)
run_test_aishell2.sh (2186, 2019-06-25)
run_test_th.sh (1526, 2019-06-25)
select_final_model.sh (247, 2019-06-25)
test.py (4364, 2019-06-25)
train.py (8419, 2019-06-25)
utils.py (12420, 2019-06-25)

近期下载者

相关文件


收藏者