Commits

Branch selector

 master

User selector

HectorSVC

Datepicker

All time

Commit History

Commits on Oct 21, 2019

• Skip GemmNoTrans_f16 test for CUDA if the hardware does not support fp16

  Show description for cdcca43
  HectorSVC

  authored
  cdcca43

  Copy full SHA for cdcca43

Commits on Oct 14, 2019

• Enable Gpu multi-device test for CUDA EP and Trt EP

  Show description for 640f71c
  HectorSVC

  authored
  640f71c

  Copy full SHA for 640f71c

Commits on Oct 11, 2019

• Fix issue that TRT not work for device other than device id 0

  Show description for 3b335c9
  HectorSVC

  authored
  3b335c9

  Copy full SHA for 3b335c9

Commits on Oct 7, 2019

• refactor Cuda Ops Sum, Max, Min, remove dup code (#1946)

  Show description for 00e24ae
  HectorSVC

  authored
  00e24ae

  Copy full SHA for 00e24ae

Commits on Sep 24, 2019

• CUDA Equal Greater Less can't support multi-directional broadcast

  Show description for e288b87
  HectorSVC

  authored
  e288b87

  Copy full SHA for e288b87

Commits on Sep 18, 2019

• remove sudo from the cleanup step for Linux so that we don't need the sudo access for vstsagent build user

  Show description for 582a27f
  HectorSVC

  authored
  582a27f

  Copy full SHA for 582a27f

Commits on Sep 12, 2019

• Fix the issue that it run into alloc failed on multiple cuda device. We have some place hard code the allocator always use device 0. (#1815)

  Show description for a0ba25f
  HectorSVC

  authored
  a0ba25f

  Copy full SHA for a0ba25f

Commits on Sep 11, 2019

• Enable Openvino nightly build on edge device (#1684)

  Show description for 2b8677b
  HectorSVC

  authored
  2b8677b

  Copy full SHA for 2b8677b

Commits on Aug 31, 2019

• Update the docker file for OpenVINO (#1741)

  Show description for dc9c895
  HectorSVC

  authored
  dc9c895

  Copy full SHA for dc9c895

Commits on Aug 29, 2019

• Fix a issue that CUDA EP fallback to much nodes to CPU for some case which cause huge data copy. If the node's inputs are all initializer, we shouldn't fallback the node to CPU. (#1727)

  Show description for 810ee00
  HectorSVC

  authored
  810ee00

  Copy full SHA for 810ee00

Commits on Aug 21, 2019

• cudnnRNNForwardInferenceEx doesn't support 0 sequence in the bathes

  Show description for e652a23
  HectorSVC

  authored
  e652a23

  Copy full SHA for e652a23

Commits on Aug 8, 2019

• Fix race condition issue in RNN/LSTM/GRU (#1544)

  Show description for 38d7854
  HectorSVC

  authored
  38d7854

  Copy full SHA for 38d7854

Commits on Aug 1, 2019

• Fix a bug in Expand cuda op implementation. (#1528)

  Show description for 57e2482
  HectorSVC

  authored
  57e2482

  Copy full SHA for 57e2482

Commits on Jul 26, 2019

• Move Class CudnnDropout to cudnn_common.h (#1492)

  Show description for 717e764
  HectorSVC

  authored
  717e764

  Copy full SHA for 717e764

Commits on Jul 23, 2019

• remove const_cast which makes it's not thread safe. (#1463)

  HectorSVC

  authored
  31838fc

  Copy full SHA for 31838fc

Commits on Jul 22, 2019

• Add test for LSTM/GRU which has shorter sequence in the middle (#1437)

  Show description for f938a6e
  HectorSVC

  authored
  f938a6e

  Copy full SHA for f938a6e

Commits on Jul 18, 2019

• CUDNN_RNN_DATA_LAYOUT_SEQ_MAJOR_UNPACKED works with CUDNN_RNN_PADDED_… (#1428)

  Show description for 1ff957f
  HectorSVC

  authored
  1ff957f

  Copy full SHA for 1ff957f

Commits on Jul 9, 2019

• Update the log message for fallback case. (#1370)

  Show description for 42c1876
  HectorSVC

  authored
  42c1876

  Copy full SHA for 42c1876

• Fix the random UT failure for RNN/GRU cases which have padded sequenc… (#1361)

  Show description for 298f305
  HectorSVC

  authored
  298f305

  Copy full SHA for 298f305

Commits on Jul 3, 2019

• Implement the Concat CUDA kernel (#1333)

  Show description for 2a6c69d
  HectorSVC

  authored
  2a6c69d

  Copy full SHA for 2a6c69d

Commits on Jun 12, 2019

• use cudnnRNNForwardInferenceEx for unpacked (padded) layout case (#899)

  Show description for bbdd1d6
  HectorSVC

  authored
  bbdd1d6

  Copy full SHA for bbdd1d6

Commits on Jun 7, 2019

• ConstantOfShape CUDA implementation (#1168)

  Show description for 8d68098
  HectorSVC

  authored
  8d68098

  Copy full SHA for 8d68098

Commits on Jun 3, 2019

• Update MaxPool & AveragePool to support opset 10 (#1141)

  Show description for b8a699f
  HectorSVC

  authored
  b8a699f

  Copy full SHA for b8a699f

Commits on May 7, 2019

• move non_max_suppression source code file to object_detection folder (#973)

  HectorSVC

  authored
  f2999cf

  Copy full SHA for f2999cf

Commits on May 3, 2019

• Add Split op CUDA implementation (#964)

  HectorSVC

  authored
  ab3355f

  Copy full SHA for ab3355f

Commits on May 1, 2019

• Add Max Cuda implementation (#950)

  Show description for 6550bb5
  HectorSVC

  authored
  6550bb5

  Copy full SHA for 6550bb5

Commits on Apr 26, 2019

• Fix a issue that ort will crash if the model has subgraph created by controlflow ops like loop, if, scan. (#917)

  Show description for 8633e9f
  HectorSVC

  authored
  8633e9f

  Copy full SHA for 8633e9f

Commits on Apr 22, 2019

• Move NMS to Onnx domain (#865)

  Show description for e8d7220
  HectorSVC

  authored
  e8d7220

  Copy full SHA for e8d7220

Commits on Apr 17, 2019

• Fix some issue in CUDA GRU and ReduceSum (#845)

  Show description for f1af493
  HectorSVC

  authored
  f1af493

  Copy full SHA for f1af493

Commits on Apr 11, 2019

• Update NMS to support max_output_boxes_per_class = 0. NMS will do nothing for this case. (#816)

  HectorSVC

  authored
  0741baf

  Copy full SHA for 0741baf

Commits on Apr 1, 2019

• Update NMS to compatible with both TF & Pytoch models (#636)

  Show description for b1115f4
  HectorSVC

  authored
  b1115f4

  Copy full SHA for b1115f4

Commits on Feb 20, 2019

• Fix a issue for cuda reduce Max Min (#474)

  Show description for 4408342
  

  HectorSVC

  authored and

  Changming Sun

  committed
  4408342

  Copy full SHA for 4408342

Commits on Feb 12, 2019

• Move build dependencies like setuptools wheel numpy into docker image (#468)

  Show description for e7c1b77
  HectorSVC

  authored
  e7c1b77

  Copy full SHA for e7c1b77

Commits on Feb 6, 2019

• Fix float 16 type support for some CUDA kernels (#436)

  Show description for f14b258
  HectorSVC

  authored
  f14b258

  Copy full SHA for f14b258

Commits on Jan 22, 2019

• use gemm to replace matmul + add (#234)

  Show description for 647cc2d
  HectorSVC

  authored
  647cc2d

  Copy full SHA for 647cc2d

Pagination

Previous

Next