1ytic warp-rnnt issues - Githubissues

1ytic / warp-rnnt

CUDA-Warp RNN-Transducer

MIT License

211 stars 41 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

__version__ assignment breaks local build.

#41 JackTemaki opened 2 months ago
0
ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory

#40 hechengzhi985 opened 5 months ago
2
Add impl of compact layout input following #26

#39 maxwellzh closed 1 year ago
3
improve efficiency of warps

#38 maxwellzh opened 1 year ago
4
Exception: CPU version is not implemented

#37 iflov opened 1 year ago
2
can't install warp-rnnt

#36 12-zhx opened 1 year ago
2
THC/THC.h: No such file or directory

#35 cchan-lm opened 1 year ago
7
rnnt_loss status 1

#34 jiay7 opened 2 years ago
2
undefined symbol: _ZNSt19basic_ostringstreamIcSt11char_traitsIcESaIcEEC1Ev

#33 wylqq312715289 opened 2 years ago
2
Strange behavior using PyTorch DDP

#32 snakers4 opened 2 years ago
7
PyTorch 1.9 Support

#31 snakers4 closed 2 years ago
5
Normalize the RNN-T Loss with input seq length

#30 aheba closed 2 years ago
1
ninja: build stopped: subcommand failed.

#29 harisgulzar1 closed 2 years ago
1
Not support for pytorch 1.7

#28 Jxu-Thu closed 2 years ago
1
Question about average_frames and reduction parmas

#27 wl-junlin opened 2 years ago
1
warp-rnnt with compact memory layout implementation

#26 maxwellzh closed 1 year ago
11
question about the gather arguments

#25 wl-junlin closed 2 years ago
1
warning that forward/backward mismatch

#24 maxwellzh opened 2 years ago
3
Implement FastEmit regularization

#23 iceychris closed 2 years ago
2
RuntimeError: rnnt_loss status 1

#22 CLC2019 closed 2 years ago
5
Transducer loss leads to memory leak

#21 YangHao97 opened 2 years ago
4
ImportError: libcudart.so.10.2: cannot open shared object file: No such file or directory

#20 skgusrb12 closed 2 years ago
3
WARNING: sample 0 [42, 26] has a forward/backward mismatch -52.543503 / 0.000000

#19 songtaoshi opened 3 years ago
3
question about the rnnt loss arguments

#18 songtaoshi opened 3 years ago
4
Exception: CPU version is not implemented

#17 yyMoming opened 3 years ago
1
[Add] Support gather log_probs inside the loss function.

#16 TeaPoly closed 3 years ago
2
Support memory efficient version run_warp_rnnt_gather.

#15 TeaPoly closed 3 years ago
0
[Add] Add tensorflow binding.

#14 TeaPoly closed 3 years ago
2
Undefined symbol when loading module

#13 mjurkus closed 3 years ago
5
RuntimeError: Trying to backward through the graph a second time, but the buffers have already been freed。

#12 lfgogogo opened 3 years ago
1
Can't install

#11 jeongHwarr closed 3 years ago
3
how to apply backward to the output of rnnt_loss

#10 huangnengCSU closed 3 years ago
0
Benchmarking methodology used is not quite correct

#9 AshishSardana closed 2 years ago
6
Not support Pytorch 1.5

#8 li563042811 closed 3 years ago
1
Not support pytorch 1.6

#7 wyattsuen closed 3 years ago
1
Issues with non-standard CUDA install

#6 StarDuster closed 3 years ago
4
Problem installation

#5 roholazandie closed 3 years ago
3
runtime error with Python 3.8

#4 funcwj closed 3 years ago
4
operating with apex?

#3 tongjinle123 opened 4 years ago
1
Not compatible with latest pytorch like 1.5.0

#1 jacob1017 closed 4 years ago
4