Hui Zhang
|
a107b75bac
|
transform; librispeech/s2 data process ok
|
3 years ago |
|
Hui Zhang
|
6569ce123d
|
lm embed and format code
|
3 years ago |
|
Hui Zhang
|
871fc5b70d
|
more utils to support kaldi/espnet data preocess
|
3 years ago |
huangyuxin
|
9e2773dffb
|
tiny change, not important
|
3 years ago |
|
huangyuxin
|
418d85ef73
|
fix some bug and complete the recog.py
|
3 years ago |
|
huangyuxin
|
e4a9328c40
|
fix some bug and complete the recog.py
|
3 years ago |
|
Hui Zhang
|
e4ecfb22fd
|
format code
|
3 years ago |
|
Hui Zhang
|
3fa2e44e89
|
more interface, trigger, extension
|
3 years ago |
|
Hui Zhang
|
c0295aa131
|
wer/cer/bleu Calculator, label smoothing func
|
3 years ago |
|
Hui Zhang
|
6a9daa800d
|
encoder.forward_one_step for prefix feat or ys
|
3 years ago |
|
Hui Zhang
|
0cd30e48a9
|
fix no_pos call
|
3 years ago |
|
Hui Zhang
|
506f2bfd20
|
add lm interface
|
3 years ago |
|
Hui Zhang
|
12ea02fc48
|
fix no pos
|
3 years ago |
|
Hui Zhang
|
e76d51fda0
|
linear embed with relu
|
3 years ago |
|
Hui Zhang
|
4f1a61011d
|
add none pos embed
|
3 years ago |
|
Hui Zhang
|
4566351127
|
add transformer lm and encoder score api
|
3 years ago |
|
Hui Zhang
|
12f788dd0e
|
Merge branch 'develop' into join_ctc
|
3 years ago |
|
Hui Zhang
|
7cfb3334e3
|
Merge pull request #927 from PaddlePaddle/nn_ctc
[asr] not change ctc grad norm manually
|
3 years ago |
|
Hui Zhang
|
dfd80b3aa2
|
recog into decoders, format code
|
3 years ago |
|
Hui Zhang
|
a4e27da64b
|
decoder with ctc prefix score
|
3 years ago |
|
Hui Zhang
|
30499a7654
|
not change ctc grad manual
|
3 years ago |
|
Hui Zhang
|
f2f305cd66
|
add recog interface
|
3 years ago |
|
huangyuxin
|
b53171694e
|
releace the <space> with ' ' in ctcdecoder
|
3 years ago |
|
huangyuxin
|
b1a90d4d7a
|
add hub for s1 in aishell and librispeech
|
3 years ago |
|
Hui Zhang
|
7402c12092
|
Merge pull request #916 from PaddlePaddle/decoder
[decoder] scorer interface and beam search
|
3 years ago |
|
Hui Zhang
|
26a4a46c41
|
fix ctc scorer name
|
3 years ago |
|
Hui Zhang
|
6257eda00e
|
add beam search with scorers
|
3 years ago |
gongel
|
5c8e66600e
|
fix: bugfix
|
3 years ago |
|
Hui Zhang
|
97136fe97f
|
trainer eval when not train
|
3 years ago |
|
Hui Zhang
|
feaf71d468
|
u2 kaldi mutli process test with batchsize one
|
3 years ago |
|
Hui Zhang
|
9e6f89bde1
|
fix score doc
|
3 years ago |
|
Hui Zhang
|
ac04615af5
|
Merge pull request #899 from Jackwaterveg/fix_bug
Revert the ctc loss to 2.1.2 and fix the bug of export model
|
3 years ago |
|
huangyuxin
|
b464e20a40
|
fix the bug of import ctc decoder
|
3 years ago |
|
huangyuxin
|
af65d5aa0c
|
revert the ctc_loss in deepspeech/__init__.py
|
3 years ago |
|
huangyuxin
|
21d6c1cee2
|
fix the bug of export model
|
3 years ago |
Jackwaterveg
|
20488c56bc
|
Merge pull request #885 from PaddlePaddle/exp
compute cmvn before build vocab
|
3 years ago |
|
Hui Zhang
|
8ebd4245d7
|
fix detoken for char
|
3 years ago |
|
huangyuxin
|
d265f9c527
|
update the wget script for openfst
|
3 years ago |
|
Hui Zhang
|
0f59459a66
|
add LayerDict
|
3 years ago |
|
Hui Zhang
|
666b42d18b
|
rename scorers
|
3 years ago |
|
Jackwaterveg
|
1a0c0ae08e
|
Merge pull request #892 from PaddlePaddle/ctc_grad_norm
[WIP] ctc loss api debug
|
3 years ago |
|
Hui Zhang
|
06a410f907
|
fix ctc_loss call for 2.2.0
|
3 years ago |
|
Hui Zhang
|
d14e56476f
|
multi workers for valid dataloader
|
3 years ago |
|
Hui Zhang
|
3c32a292e9
|
Merge pull request #889 from PaddlePaddle/parse_file
argparse load conf from file
|
3 years ago |
|
Hui Zhang
|
ffd34bac6c
|
Merge pull request #890 from PaddlePaddle/format
format code
|
3 years ago |
|
Hui Zhang
|
10a2da6808
|
format code
|
3 years ago |
|
Hui Zhang
|
37790b76d0
|
argparse load conf from file
|
3 years ago |
|
Hui Zhang
|
4cc6a18fa6
|
rm useless swig dir
|
3 years ago |
|
Hui Zhang
|
169823ce0b
|
remove useless device
|
3 years ago |
|
Hui Zhang
|
54b31d35f1
|
decoder end detection, when no more hyps score larger than before by a margin meet M count
|
3 years ago |
