Issues
- 2
Exception: CPU version is not implemented
#37 opened by iflov - 0
__version__ assignment breaks local build.
#41 opened by JackTemaki - 2
ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory
#40 opened by hechengzhi985 - 2
undefined symbol: _ZNSt19basic_ostringstreamIcSt11char_traitsIcESaIcEEC1Ev
#33 opened by wylqq312715289 - 2
rnnt_loss status 1
#34 opened by jiay7 - 2
can't install warp-rnnt
#36 opened by 12-zhx - 4
runtime error with Python 3.8
#4 opened by funcwj - 4
improve efficiency of warps
#38 opened by maxwellzh - 7
THC/THC.h: No such file or directory
#35 opened by cchan-lm - 7
Strange behavior using PyTorch DDP
#32 opened by snakers4 - 3
WARNING: sample 0 [42, 26] has a forward/backward mismatch -52.543503 / 0.000000
#19 opened by songtaoshi - 4
question about the rnnt loss arguments
#18 opened by songtaoshi - 5
RuntimeError: rnnt_loss status 1
#22 opened by CLC2019 - 5
PyTorch 1.9 Support
#31 opened by snakers4 - 1
Normalize the RNN-T Loss with input seq length
#30 opened by aheba - 1
Not support for pytorch 1.7
#28 opened by Jxu-Thu - 1
ninja: build stopped: subcommand failed.
#29 opened by harisgulzar1 - 3
ImportError: libcudart.so.10.2: cannot open shared object file: No such file or directory
#20 opened by skgusrb12 - 1
- 1
question about the gather arguments
#25 opened by wl-junlin - 6
- 3
warning that forward/backward mismatch
#24 opened by maxwellzh - 4
Transducer loss leads to memory leak
#21 opened by YangHao97 - 1
RuntimeError: Trying to backward through the graph a second time, but the buffers have already been freed。
#12 opened by lfgogogo - 1
Exception: CPU version is not implemented
#17 opened by yyMoming - 5
Undefined symbol when loading module
#13 opened by mjurkus - 3
Can't install
#11 opened by jeongHwarr - 0
how to apply backward to the output of rnnt_loss
#10 opened by huangnengCSU - 3
Problem installation
#5 opened by roholazandie - 1
Not support Pytorch 1.5
#8 opened by li563042811 - 1
Not support pytorch 1.6
#7 opened by wyattsuen - 4
Issues with non-standard CUDA install
#6 opened by StarDuster - 4
- 1
operating with apex?
#3 opened by tongjinle123