8000 Commits · Zhangjh111/kaldi · GitHub

More Web Proxy on the site http://driver.im/

Commits

Commits on Nov 1, 2019

[src] cuda batched decoder, fix memory bugs (#3697)

luitjens
authored and
danpovey
committed

Commits on Oct 19, 2019

[src] CUDA decoder: write all output to single stream (#3666)

luitjens
authored and
danpovey
committed

Commits on Oct 17, 2019

[src] CUDA batched decoder pipeline fixes (#3659)

luitjens
authored and
danpovey
committed
[src] Some CUDA i-vector fixes (#3660)

luitjens
authored and
danpovey
committed

Commits on Sep 19, 2019

[src] Fix two bugs in batched-wav-nnet3-cuda binary. (#3600)

luitjens
authored and
danpovey
committed

Commits on Jul 2, 2019

[src] Marking operator as __host__ __device__ to avoid build issues (#3441)

luitjens
authored and
danpovey
committed

Commits on Jun 26, 2019

[src] Enable an option to use the GPU for feature extraction in GPU decoding (#3420)

luitjens
authored and
danpovey
committed
[src] Fix build of online feature extraction with older CUDA version (#3415)

luitjens
authored and
danpovey
committed

Commits on Jun 25, 2019

[src] Added error string to CUDA allocation errors. (#3413)

luitjens
authored and
danpovey
committed

Commits on Jun 24, 2019

[src] Fix build error introducted by race condition in PR requests/accepts. (#3412)

luitjens
authored and
danpovey
committed
[src] Added GPU feature extraction (will improve speed of GPU decoding) (#3390)

luitjens
authored and
danpovey
committed
[src] Remove debugging assert in cuda feature extraction code (#3411)

luitjens
authored and
danpovey
committed

Commits on Jun 11, 2019

[src] Implemented CUDA acclerated online cmvn. (#3370)

luitjens
authored and
danpovey
committed
[build,src] Enhancements to the cudamatrix/cudavector classes. (#3373)

luitjens
authored and
danpovey
committed

Commits on Jun 6, 2019

[src] Batch nnet3 optimizations. Batch some of the copies in and copies out (#3378)

luitjens
authored and
danpovey
committed
[src] Fix non-cuda `make depend` build by putting compile guards around header. (#3374)

luitjens
authored and
danpovey
committed

Commits on Jun 3, 2019

[src] Optimizations for batch nnet3. The issue fixed here is that (#3351)

luitjens
authored and
danpovey
committed
[src] Add CUDA accelerated MFCC computation. (#3348)

luitjens
authored and
danpovey
committed

Commits on May 23, 2019

[src] Add interfaces to nnet-batch-compute that expects device input. (#3311)

luitjens
authored and
danpovey
committed

Commits on Apr 25, 2019

[src] Fixes to cuda unit tests. (#3268)

luitjens
authored and
danpovey
committed

Commits on Apr 22, 2019

[src] GPU-related changes for speed and correctness on newer arch's. (#3221)

luitjens
authored and
danpovey
committed

Commits on Apr 7, 2019

[src] Move curand handle out of CuRand class and into CuDevice. (#3196)

luitjens
authored and
danpovey
committed

Commits on Mar 30, 2019

[src] Efficiency improvement and extra checking for cudamarix, RE default stream (#3182)

luitjens
authored and
danpovey
committed

Commits on Mar 27, 2019

[build] Don't build for Tegra sm_XX versions on x86/ppc and vice versa; allow --cuda-arch overrides to have multiple versions (#3171)

luitjens
authored and
danpovey
committed

Commits on Mar 4, 2019

[src] Add device options to enable tensor core math mode. (#3066)

luitjens
authored and
danpovey
committed
[src] Add GPU version of MergeTaskOutput().. relates to batch decoding (#3067)

luitjens
authored and
danpovey
committed
[src] Make copies occur in per-thread default stream (for GPUs) (#3068)

luitjens
authored and
danpovey
committed

0