HawkAaron/warp-transducer issues and pull requests

#102 - Impossible to compile with CUDA in an Anaconda env

Issue - State: closed - Opened by TParcollet over 1 year ago - 1 comment

#101 - Update c++17 for pytorch 2.1.0

Pull Request - State: open - Opened by yanbing-j over 1 year ago - 6 comments

#100 - rnnt_loss compute bptt?

Issue - State: open - Opened by YooSungHyun over 2 years ago

#99 - "python setup.py install" error: command '/usr/bin/gcc' failed with exit code 1

Issue - State: open - Opened by Schwein3211 over 2 years ago - 1 comment

#98 - make fails on A40 with CUDA 11.3

Issue - State: open - Opened by jiay7 almost 3 years ago - 2 comments

#97 - nvcc fatal : Unsupported gpu architecture 'compute_30' while execute make again

Issue - State: closed - Opened by Stasiche about 3 years ago - 5 comments

#96 - There is a problem about training a RNN-T model?

Issue - State: open - Opened by scufan1990 about 3 years ago - 4 comments

#95 - segfault with large tensor for pytorch

Issue - State: open - Opened by Peach-He about 3 years ago - 1 comment

#94 - Can someone explain how the memory is allocated for alpha, beta and log_probs2?

Issue - State: open - Opened by ronggong about 3 years ago

#93 - the grad of cpu ,cuda are not same

Issue - State: open - Opened by lcylcy over 3 years ago - 1 comment

#92 - RNNT loss is constantly zero

Issue - State: open - Opened by Rajratnpranesh over 3 years ago - 1 comment

#91 - building error

Issue - State: open - Opened by jinggaizi over 3 years ago - 1 comment

#90 - Update pytorch_binding/setup.py

Pull Request - State: closed - Opened by JinuJeong almost 4 years ago

#89 - test_gpu fails on RTX 3090 with CUDA 11.1

Issue - State: open - Opened by mdabek almost 4 years ago - 2 comments

#88 - Getting segmentation fault on CentOS

Issue - State: open - Opened by stefan-falk almost 4 years ago

#87 - GPU execution requested, but not compiled with GPU support

Issue - State: open - Opened by jungwook518 about 4 years ago - 1 comment

#86 - Negative loss in test with multiple batches on GPU

Issue - State: open - Opened by gandroz about 4 years ago

#85 - cuda_curand_library not found at build time

Issue - State: closed - Opened by gandroz about 4 years ago

#84 - Correcting two small typos in messages

Pull Request - State: open - Opened by vincentqb about 4 years ago

#83 - Update pytorch data pointer

Pull Request - State: open - Opened by vincentqb about 4 years ago

#82 - enable static build

Pull Request - State: closed - Opened by vincentqb about 4 years ago - 1 comment

#81 - AttributeError: module 'warprnnt_pytorch.warp_rnnt' has no attribute 'gpu_rnnt'

Issue - State: open - Opened by li563042811 about 4 years ago - 1 comment

#80 - ModuleNotFoundError in Google Colab with GPU runtime

Issue - State: closed - Opened by vincentqb about 4 years ago - 1 comment

#79 - nvcc fatal : Unsupported gpu architecture 'compute_30' while execute make

Issue - State: closed - Opened by GGzhangBOY about 4 years ago - 2 comments

#78 - undefined symbol issue when importing TensorFlow binding

Issue - State: closed - Opened by naxingyu about 4 years ago - 1 comment

#77 - AttributeError: module 'tensorflow.python.framework.ops' has no attribute 'RegisterShape'

Issue - State: open - Opened by mayankpathaklumiq over 4 years ago - 2 comments

#76 - Fixed CUDA error

Pull Request - State: open - Opened by ncilfone over 4 years ago - 1 comment

#75 - "import warprnnt_tensorflow" got "undefined symbol: _ZN10tensorflow*" error

Issue - State: open - Opened by huangzj421 over 4 years ago - 2 comments

#74 - question about w

Issue - State: open - Opened by farmboy-dev over 4 years ago

#73 - building problem

Issue - State: closed - Opened by rhoposit over 4 years ago - 1 comment

#72 - rnnt_loss() gives different outputs (increasing) executed on same input

Issue - State: open - Opened by stefan-falk over 4 years ago - 1 comment

#71 - Some questions about RegisterShape

Issue - State: open - Opened by alongwithyou over 4 years ago - 5 comments

#70 - fatal error: cuda_runtime_api.h: No such file or directory

Issue - State: open - Opened by MorrisXu-Driving over 4 years ago

#69 - CPU gradient definitions

Issue - State: open - Opened by xkr1 almost 5 years ago - 3 comments

#68 - TF2.0 test_multiple_batches_gpu failed due to negative loss

Issue - State: open - Opened by tvatsal1996 almost 5 years ago

#67 - import error when use warp_rnnt

Issue - State: open - Opened by shunfeichen almost 5 years ago - 2 comments

#66 - Support pytorch1.5

Pull Request - State: closed - Opened by kamo-naoyuki almost 5 years ago - 1 comment

#65 - Unknown shaped and negative loss

Issue - State: closed - Opened by tvatsal1996 almost 5 years ago - 5 comments

#64 - Fix illegal memory access of compute_grad_kernel

Pull Request - State: closed - Opened by jaesong almost 5 years ago - 1 comment

#63 - test.py error

Issue - State: closed - Opened by BuaaAlban almost 5 years ago - 4 comments

#62 - cant find file "tensorflow/core/framework/op.h

Issue - State: closed - Opened by ybNo1 almost 5 years ago - 3 comments

#61 - CUDA error: an illegal memory access was encountered

Issue - State: open - Opened by FactoDeepLearning almost 5 years ago - 12 comments

#60 - segfault for pytorch

Issue - State: closed - Opened by YimengZhu almost 5 years ago - 9 comments

#59 - error in tf2.1

Issue - State: open - Opened by nichongjia-2007 about 5 years ago - 4 comments

#58 - cmake problem

Issue - State: closed - Opened by yjiangling about 5 years ago - 4 comments

#57 - add_network

Issue - State: closed - Opened by lijiangguo about 5 years ago - 1 comment

#56 - fatal error: tensorflow/core/framework/bounds_check.h: No such file or directory #include "tensorflow/core/framework/bounds_check.h"

Issue - State: closed - Opened by yjiangling about 5 years ago - 6 comments

#55 - question about gradient to last blank transition

Issue - State: closed - Opened by swpd about 5 years ago - 1 comment

#54 - multi-gpu

Issue - State: closed - Opened by lijiangguo over 5 years ago - 3 comments

#53 - Can blank assign with non-zero?

Issue - State: closed - Opened by yanz0920 over 5 years ago - 1 comment

#52 - TF 2.0 Support

Issue - State: closed - Opened by dtchantao over 5 years ago - 1 comment

#51 - Adding rnnt_loss to Keras ("Cannot iterate over a shape with unknown rank.")

Issue - State: closed - Opened by sephiroce over 5 years ago - 2 comments

#50 - negative loss

Issue - State: closed - Opened by huangnengCSU over 5 years ago - 12 comments

#49 - How can I build the pkg in windows?

Issue - State: closed - Opened by Times125 over 5 years ago - 2 comments

#48 - Error during building warp-transducer: __shfl_down() is not defined

Issue - State: closed - Opened by AbrarAbdulnabi91 over 5 years ago - 1 comment

#47 - tensorflow setup on CPU error

Issue - State: closed - Opened by zhoudoufu over 5 years ago - 2 comments

#46 - Update setup.py

Pull Request - State: closed - Opened by PoDuReM over 5 years ago

#45 - Update setup.py

Pull Request - State: closed - Opened by PoDuReM over 5 years ago

#44 - dynamic module does not define module export function(PyInit_warp_rnnt)

Issue - State: closed - Opened by shunfeichen over 5 years ago - 2 comments

#43 - Remove the union from rnntOptions to solve segmentation fault issues

Pull Request - State: closed - Opened by jannepy over 5 years ago

#42 - Can log_probs be torch.float16?

Issue - State: closed - Opened by Marcovaldong over 5 years ago - 4 comments
Labels: enhancement, help wanted

#41 - question about input shape

Issue - State: closed - Opened by huangnengCSU over 5 years ago - 2 comments

#40 - Could you please say what can be improved in this implementation?

Issue - State: closed - Opened by PoDuReM over 5 years ago

#39 - still pytorch binding GPU test error

Issue - State: closed - Opened by qmpzzpmq over 5 years ago - 1 comment

#38 - Added support for tf version >= 1.14

Pull Request - State: closed - Opened by IsaacTay over 5 years ago

#37 - error in make

Issue - State: closed - Opened by jackson1895 over 5 years ago - 5 comments

#36 - AttributeError: module 'warprnnt_pytorch' has no attribute 'cpu_rnnt

Issue - State: closed - Opened by germany-zhu over 5 years ago - 6 comments

#35 - import warprnnt_pytorch as warp_rnnt

Issue - State: closed - Opened by germany-zhu over 5 years ago

#34 - getting negative loss

Issue - State: closed - Opened by ni1lloc over 5 years ago - 2 comments

#33 - ImportError: libwarprnnt.so: cannot open shared object file: No such file or directory

Issue - State: closed - Opened by zw76859420 over 5 years ago - 2 comments

#32 - Update setup.py

Pull Request - State: closed - Opened by FieldRen over 5 years ago - 1 comment

#31 - tensorflow binding undefined symbol: _ZTIN10tensorflow8OpKernelE

Issue - State: closed - Opened by Slyne over 5 years ago - 1 comment

#30 - tensorflow test failed at test_multiple_batches_gpu with results dismatch

Issue - State: closed - Opened by qppp558 over 5 years ago - 4 comments

#29 - GPU test segmentation fault with CUDA 10.0

Issue - State: closed - Opened by sephiroce over 5 years ago - 11 comments
Labels: help wanted

#28 - GPU Test Failed

Issue - State: closed - Opened by rajeevbaalwan over 5 years ago - 24 comments

#27 - RNNTloss decrease while the syllabel error rate increace

Issue - State: closed - Opened by liangzhenlin123 over 5 years ago - 14 comments

#26 - unable to run test.py file

Issue - State: closed - Opened by rajeevbaalwan over 5 years ago - 4 comments

#25 - What's the license of this project

Issue - State: closed - Opened by ling0322 over 5 years ago - 1 comment

#24 - tf-binding Registering two gradient with name 'WarpRNNT' !

Issue - State: closed - Opened by suenyandan almost 6 years ago - 9 comments

#23 - PyTorch reduction='none' is not properly handled

Issue - State: closed - Opened by chanil1218 almost 6 years ago - 1 comment

#22 - PyTorch non-static autograd.Function loss cause issue

Issue - State: closed - Opened by chanil1218 almost 6 years ago - 4 comments

#21 - negative loss in tf version

Issue - State: closed - Opened by imodile almost 6 years ago - 4 comments

#20 - Program test cases do not pass

Issue - State: closed - Opened by liqiang4113 almost 6 years ago - 7 comments

#19 - Can i use RNNTLoss in transducer_np.py while training transducer-model?

Issue - State: closed - Opened by xingchensong almost 6 years ago - 6 comments

#18 - does this binding support multi-gpus?

Issue - State: closed - Opened by Marcovaldong almost 6 years ago - 2 comments

#17 - GPU version gives nan value

Issue - State: closed - Opened by adonisues almost 6 years ago - 9 comments

#16 - Grads mismatch when running tests

Issue - State: closed - Opened by ni1lloc almost 6 years ago - 2 comments

#15 - fatal error: cuda_runtime_api.h: No such file or directory

Issue - State: closed - Opened by imr555 almost 6 years ago - 20 comments

#14 - Negative value of loss function

Issue - State: closed - Opened by ZhengkunTian about 6 years ago - 20 comments

#13 - pytorch binding for pytorch 1.0

Issue - State: closed - Opened by ybzhou over 6 years ago - 3 comments

#12 - joint network defined in source code

Issue - State: closed - Opened by rzcwade over 6 years ago - 4 comments

#11 - tensorflow binding

Issue - State: closed - Opened by rzcwade over 6 years ago - 2 comments

#10 - RNNT Loss tensorflow binding

Issue - State: closed - Opened by rzcwade over 6 years ago

#9 - Input activation format

Issue - State: closed - Opened by ybzhou over 6 years ago - 2 comments

#8 - warp rnnt tensorflow binding issue

Issue - State: closed - Opened by rzcwade over 6 years ago - 6 comments

#7 - cuda runtime error (77) : an illegal memory access was encountered

Issue - State: closed - Opened by simonwang517 over 6 years ago - 10 comments

#6 - build against pip-installed tensorflow-gpu gives segfault

Issue - State: closed - Opened by fginter over 6 years ago - 5 comments

GitHub / HawkAaron/warp-transducer issues and pull requests