285 Commits (a64920c46802a3030bfdf8754870ffb1e5c6ba78)

Author SHA1 Message Date
  wangshuide2020 6cbe8dd02e optimizes the kernel error description of LSTM, Pad, ReLU, etc. 4 years ago
  wangshuide2020 674e3aa9d6 optimizes the kernel error description of Adagrad, Adam, Conv2d, etc. 4 years ago
  zhunaipan 8ce4e62725 optimize the comment and log description 4 years ago
  wilfChen 54761ecccc codedex 4 years ago
  zhangyihui a94b3dbcfe clean up the static alarms of the second batch of operator groups 4 years ago
  i-robot 4c8854ac02 !24403 Clean up the static alarms of the first batch of operator groups 4 years ago
  i-robot 2c1d3baace !24269 add vector size check, input shape check and divide by zero check for gpu operators. 4 years ago
  markuskunej abdba421e5 added GetReductionInt to common_utils.h and replaced duplicated code in all loss with reduction gpu op kernels (nll loss, kl div loss, and binary cross entropy) 4 years ago
  zhangyihui 27a80a75c0 Clean up the first batch of static alarms of operator group 4 years ago
  wangshuide2020 7a1862a6e6 add vector size check, input shape check and divide by zero check for gpu operators. 4 years ago
  i-robot e81476d1b0 !24257 erase warning 4 years ago
  zong_shuai 417096ddcc erase_warning_1 4 years ago
  wangshuide2020 a35a1fe67d add vector size check, nullptr check and clean code for gpu operators. 4 years ago
  wangshuide2020 e06beb2ed4 add validation of vector size and non-zero validation of denominator for nn gpu operators. 4 years ago
  liangchenghui 7614129eda Add isnan/isinf operator bprop. 4 years ago
  i-robot b49174cd40 !23096 Generalize gpu PadOp to support more than 4 dimensions 4 years ago
  Peilin Wang d7b23ca4b8 fix pad 4 years ago
  i-robot 63114a3dfd !22385 Fix bug in GPU conv3dtranspose 4 years ago
  simson 8a0087bceb fix precision error of resizebilineargrad 4 years ago
  fan1997 84a540e743 Fix bug in conv3dtranspose gpu 4 years ago
  i-robot f536d88570 !22821 modify resizebilineargrad input type 4 years ago
  simson f00e22342b modify resizebilineargrad input type 4 years ago
  i-robot 5e6287bec1 !22585 Add grad implementation of AdaptiveAvgPool2D 4 years ago
  zuochuanyong 068191f222 add AdaptiveAvgPool2DGrad op 4 years ago
  simson 7a2fbdda85 modify resizebilinear infer type 4 years ago
  djc b077aa1cab [feat] [assistant] [I3T96T] add new Dataset operator CMUARCTICDataset 4 years ago
  djc 4e6f7dc97d [feat] [assistant] [I3T96X] add new Dataset operator LibriSpeechDataset 4 years ago
  Peilin Wang 6a1b1495d9 initial commit: add nullptr exception in GetDeviceAddress 4 years ago
  i-robot 22e9299c17 !20885 add dtypes & fft kernels for SPONGE 4 years ago
  huangmengxi e32297dc6b add dtypes for sponge 4 years ago
  danishfarid 92d9bc7ccd fix for async mem_init bilinearResize_grad 4 years ago
  Peilin Wang 594571fd4c initial commit: fix 11 dts tickets 4 years ago
  zuochuanyong 1d565f9f8a support MaxPool3DGrad on GPU 4 years ago
  i-robot d16e9bc3f6 !20369 GPU fix maxpoolgrad 4 years ago
  i-robot fb33ba2b47 !19941 [MS][GPU] resizeBilinearGrad - Op FP16 fix 4 years ago
  VectorSL a3590bca46 fix maxpool grad 4 years ago
  danishfarid aa37923aa5 first commit 4 years ago
  buxue 2b2efb0a75 fix prelu weight grad accuracy error fp16 on GPU 4 years ago
  buxue 5bf41bfbd2 improve PReLU forward and implement backward on GPU 4 years ago
  i-robot 8e043090be !18472 Implement UNet3d on GPU 4 years ago
  likesen 99a995b432 Implement UNet3d on GPU 4 years ago
  zuochuanyong e890c2a2ae fix accuracy error when input H is not equal to W 4 years ago
  i-robot 26c7d274c9 !18441 Fix conv3d cudnn algorithm error 4 years ago
  i-robot 6c33e0b710 !18392 fix the exception when occur error and replace magic number with const value. 4 years ago
  i-robot 9085be08b9 !18163 Support ConvTranspose3D on GPU 4 years ago
  tom__chen 35f6a1af56 fix conv3d cudnn algorithm error 4 years ago
  Li Kesen 7d94095730 Support Conv3dTranspose for GPU 4 years ago
  wangshuide2020 30690b1f27 fix the exception when occur error and replace magic number with const value. 4 years ago
  buxue d50d46013b code security check 4 years ago
  markuskunej 2fece8a7c2 added nll_loss_grad for gpu 4 years ago