Commit Graph

  • d06d7054b4 [v1.x] More onnx export updates (#19692) Joe Evans 2020-12-17 21:45:53 -08:00
  • 7baac50850 [v1.x] onnx test coverage for leakyrelu elemwise_add concat activation (#19687) Zhaoqi Zhu 2020-12-17 19:57:34 -08:00
  • 94cbc88139 [v1.x] Add more onnx export unit tests, refactor onnxruntime tests. (#19689) Joe Evans 2020-12-17 13:55:53 -08:00
  • c53e32b953 [v1.x] Onnx fix slice_axis and embedding and reshape (#19677) Zhaoqi Zhu 2020-12-17 10:24:52 -08:00
  • 4e109ce855 [v1.x] Add coverage to onnx test pipeline. (#19682) Joe Evans 2020-12-16 21:21:36 -08:00
  • f2ab4c9db8 ONNX _contrib_interleaved_matmul_selfatt_valatt and LayerNorm (#19661) waytrue17 2020-12-16 17:22:08 -08:00
  • 6b65b91b04 [v1.x] Update onnx export support for FullyConnected and add unit tests (#19679) Joe Evans 2020-12-16 10:24:50 -08:00
  • 91503f71e1 fix typo in adamw (#19680) Zheyu Ye 2020-12-16 14:59:20 +08:00
  • 9ccc53dd27 large tensor tests batch 6 (#19617) Zhaoqi Zhu 2020-12-15 22:15:33 -08:00
  • 806d91f20e [v1.x] Make ONNX export operators work properly with the node input shape (#19676) Joe Evans 2020-12-15 22:12:45 -08:00
  • b3d9af2351 onnx support more ops (#19653) Zhaoqi Zhu 2020-12-15 11:07:07 -08:00
  • 2194c0c532 fixed macros with name (#19669) Sam Skalicky 2020-12-15 10:12:58 -08:00
  • 57478a6087 fixed macros with name (#19668) Sam Skalicky 2020-12-15 10:12:50 -08:00
  • b5fd18a139 [v1.x] Save/Load Gluon Blocks & HybridBlocks (#19565) Sam Skalicky 2020-12-15 10:12:40 -08:00
  • f3455f72af Fix local variable 'optimizer' referenced before assignment (#19666) Harshit Sharma 2020-12-15 09:11:54 -08:00
  • 05fb317114 [submodule] Upgrade oneDNN to v2.0 (#19667) bartekkuncer 2020-12-15 16:57:16 +01:00
  • a87d3f41fa Improve onnx test suite (#19662) Zhaoqi Zhu 2020-12-14 12:47:16 -08:00
  • 765297bf5d Save/Load Gluon Blocks & HybridBlocks (#19564) Sam Skalicky 2020-12-14 11:06:39 -08:00
  • ed4aa23dba Change __str__ representation of bfloat16 NDArray from uints to floats (#19608) Paweł Głomski 2020-12-14 19:09:28 +01:00
  • 3746babc8f [FEATURE] Restore Quantization API to MXNet (#19587) bgawrych 2020-12-12 05:31:00 +01:00
  • 87674ea21e NP Large tensor tests batch 5 (#19541) Zhaoqi Zhu 2020-12-11 10:58:49 -08:00
  • b219ac2e72 [v1.x] Add more ONNX export support to operators (#19625) Joe Evans 2020-12-11 10:22:19 -08:00
  • d94e47c8e3 [v1.8.x] Backport #19656 - fix R builds (#19658) Joe Evans 2020-12-11 07:53:26 -08:00
  • c37d5aaccf Fix R builds on CI (#19656) Joe Evans 2020-12-10 21:50:41 -08:00
  • bfbfe3f4a0 Numpy large tensor test batch 4 (#19294) Zhaoqi Zhu 2020-12-10 12:36:16 -08:00
  • 63f2b3213f [BUGFIX] Fix backward pass for nested CachedOp (#19614) waytrue17 2020-12-09 22:33:36 -08:00
  • 99e94ec89d Add tests for DL layers with large inputs (#19612) Rohit Kumar Srivastava 2020-12-09 15:10:25 -08:00
  • 1f076931f5 Enable large tensor inputs for Deconvolution layer (#19626) Rohit Kumar Srivastava 2020-12-08 18:23:29 -08:00
  • 6cbd3ae772 [DOC] Fix warnings in tutorials and turn on -W (#19624) Sheng Zha 2020-12-08 11:13:09 -05:00
  • c9841a8677 BUILD_CYTHON_MODULES=ON by default leezu-patch-2 Leonard Lausen 2020-12-08 07:06:42 -08:00
  • f8d3c14a70 update docker image used in master nightly tests Zhaoqi Zhu 2020-12-07 22:50:55 -08:00
  • bbfccaebb7 Change gcc 8 PPA to ppa:jonathonf/gcc (#19638) kpuatamazon 2020-12-08 01:55:21 +00:00
  • 4405906977 Tests for convolution layer with large inputs (#19271) Rohit Kumar Srivastava 2020-12-07 16:40:33 -08:00
  • 21bd9c776a Fix formatting of MXNet crash course material (#19629) Vidya Sagar Ravipati 2020-12-07 16:02:53 -08:00
  • 1c2a83b5e5 Fix CI README unittest_ubuntu_python3_armv7 -> unittest_ubuntu_python3_arm (#19637) kpuatamazon 2020-12-07 23:59:20 +00:00
  • 4879c7a8ad Forward port Support destructors for custom stateful ops (#19607) (#19632) Sam Skalicky 2020-12-07 15:58:56 -08:00
  • 35cab31c89 Support destructors for custom stateful ops (#19607) Sam Skalicky 2020-12-04 13:35:46 -08:00
  • b1b50210f3 Support static link to openblas & autodetect LP64 vs ILP64 settings (#19174) Zhaoqi Zhu 2020-12-04 11:00:18 -08:00
  • 2a20916f2f Skip test_sparse_mathematical_core (#19627) Leonard Lausen 2020-12-04 10:01:18 -07:00
  • 87d69faa8f Masked softmax (& log-softmax) (#19460) Moises Hernandez 2020-12-03 20:12:25 -08:00
  • f9a6369714 Fix MKL_USE_STATIC_LIBS=0 and test on one CI pipeline leezu-patch-1 Leonard Lausen 2020-12-01 16:32:28 +00:00
  • afc76b0f82 Update Ubuntu images used on CI to 20.04 (#19588) Leonard Lausen 2020-12-03 19:58:14 -07:00
  • d348e5d447 Remove obsolete six dependency (#19620) Ziyi Mu 2020-12-03 07:51:34 -08:00
  • 619eab2657 Numpy unique repeat indices large tensor checks (#19382) Zhaoqi Zhu 2020-12-02 17:18:06 -08:00
  • e6a855c476 Fix flaky onnxruntime tests. (#19618) Joe Evans 2020-12-02 16:56:04 -08:00
  • 29dd4813bc Remove obsolete six dependency (#19604) Ziyi Mu 2020-12-02 14:31:39 -08:00
  • af26e92c5c Update TVM integration to v0.7 (#19613) Leonard Lausen 2020-12-02 11:23:52 -07:00
  • 11a7903c09 run lt nightly tests as individual sub-process (#19576) Rohit Kumar Srivastava 2020-12-01 12:35:23 -08:00
  • 78fc40f68e Fix serialization bug for writing large arrays to npz (#19596) Leonard Lausen 2020-11-30 16:10:35 -07:00
  • 07bfe55f82 parametrize test_deformable_convolution (#19600) Leonard Lausen 2020-11-30 13:47:40 -07:00
  • 2db601a5bb Mark test_cv_model_inference_onnxruntime_mobilenetv2_1_0 as flaky (#19590) Leonard Lausen 2020-11-28 18:50:14 -07:00
  • 748e100929 [FEATURE] Add BN+ReLU -> BatchNormWithRelu fuse to MKLDNN backend (#19578) Adam 2020-11-29 02:19:57 +01:00
  • fef85c1462 Remove einsum overhead (#19421) MoisesHer 2020-11-27 15:19:58 -08:00
  • 36b54b94e6 Switch npx.savez from libzip to miniz (#19584) Leonard Lausen 2020-11-26 21:08:35 -07:00
  • d599a15342 Update cmake/BuildTVM.cmake to use system llvm (#19589) Leonard Lausen 2020-11-26 09:51:48 -07:00
  • 16be29ded6 Mark test_dropout as flaky (#19553) Leonard Lausen 2020-11-25 15:32:57 -07:00
  • f97544e6cb Update ONNX support to 1.7 (#19573) Joe Evans 2020-11-24 13:52:02 -08:00
  • 2ad91d2bcb [1.8.x][submodule] Upgrade to oneDNN v1.7 (#19560) bartekkuncer 2020-11-24 14:11:38 +01:00
  • 78a0f7bba4 [submodule] Upgrade to oneDNN v1.7 (#19558) bartekkuncer 2020-11-24 13:25:59 +01:00
  • 221344e282 Convert symbol to numpy symbol in Symbol class (#19523) bgawrych 2020-11-23 19:33:20 +01:00
  • 13e9d572b3 Fix compilation with USE_LAPACK=0 USE_BLAS=mkl (#19570) Andrzej Kotłowski 2020-11-21 00:07:46 +01:00
  • 66488662ea Add AMP patching of npi ops in _api_internal module (#19488) mk-61 2020-11-19 15:58:32 -08:00
  • 5dc404dfb0 NumPy compatible serialization API (#19417) Leonard Lausen 2020-11-19 15:44:37 -07:00
  • c9d9c85d19 Mark test_np_around as flaky (#19554) Leonard Lausen 2020-11-19 12:09:30 -07:00
  • db080058fd [1.x][submodule] Upgrade to oneDNN v1.7 (#19559) bartekkuncer 2020-11-19 17:52:07 +01:00
  • 7df16cba2f remove cleanup on side threads (#19557) waytrue17 2020-11-18 16:08:10 -08:00
  • dbcd6a19f1 Separate backend from hybridize and refactor optimize_for kwargs (#19543) Serge Panev 2020-11-18 15:03:31 -08:00
  • 6fa208b4ae Fix compilation for large tensor with MKL (#19067) Andrzej Kotłowski 2020-11-18 21:45:59 +01:00
  • f998ac6299 Mark test_np_average as flaky (#19538) Leonard Lausen 2020-11-17 19:56:28 -07:00
  • d2818b46ec CPU reduce kernel optimization (#19419) Zhaoqi Zhu 2020-11-17 21:17:43 -05:00
  • 18a784a327 Numpy take large tensor fix (#19325) Zhaoqi Zhu 2020-11-17 19:43:26 -05:00
  • 07c1e1e590 Don't declare/use CheckPvalGPU on CPU builds (#19545) Leonard Lausen 2020-11-17 16:41:18 -07:00
  • d2600f21c2 [1.x] Backport Faster pointwise fusion graph pass (#19269) (#19413) Przemyslaw Tredak 2020-11-17 15:26:17 -08:00
  • ab7ad4df3b Add AdamW optimizer (#19526) xinyual 2020-11-18 04:34:05 +08:00
  • fcfef81c36 Enable Large Tensor Support by default (#18625) Rohit Kumar Srivastava 2020-11-17 12:09:42 -08:00
  • 5000ccbafe Update pip install instruction (#19549) Zhaoqi Zhu 2020-11-17 12:29:50 -05:00
  • 9625f8fa74 Fix fallback.py and test_numpy_interoperability.py (#19457) barry-jin 2020-11-16 21:33:10 -08:00
  • d0710e2a01 [1.x] Move block.optimize_for backend_opts to kwargs (#19386) Serge Panev 2020-11-16 09:53:31 -08:00
  • 2e86b92c61 Don't use namespace for pow() function, since it is built into cuda math library, and cast the second argument so it will find an acceptable form. (#19533) Joe Evans 2020-11-14 11:00:00 -08:00
  • 1f15d98f98 Don't use namespace for pow() function, since it is built into cuda math library, and cast the second argument so it will find an acceptable form. (#19532) Joe Evans 2020-11-14 10:59:45 -08:00
  • 6bc064771d External Operators 2 (#19431) Sam Skalicky 2020-11-13 23:13:06 -08:00
  • 3bf556dedc Cuda 11 build fixes (#19530) Joe Evans 2020-11-13 22:45:56 -08:00
  • bb8f6cedb3 change partitioning default to False for hybridize and optimize_for (#19531) Manu Seth 2020-11-13 20:08:31 -08:00
  • d43f439e46 Enable Numpy Large tensor nightly tests (#19529) Rohit Kumar Srivastava 2020-11-13 19:47:21 -08:00
  • 4da8783c44 Fix uninitialized variable in RTC path (#19528) Przemyslaw Tredak 2020-11-13 15:59:16 -08:00
  • 6ca77ddabd [v1.x] Backport extension bug fixes to v1.x (#19469) (#19503) Sam Skalicky 2020-11-12 16:04:58 -08:00
  • 272151883f [v1.8.x] Backport extension bug fixes to v1.8.x (#19469) (#19504) Sam Skalicky 2020-11-12 16:04:52 -08:00
  • abeaba7804 disable graph partition by default (#19525) Manu Seth 2020-11-12 15:59:46 -08:00
  • f1ed842473 update cudnn from 7 to 8 for cu102 (#19522) Manu Seth 2020-11-12 13:53:30 -08:00
  • ea4e2b3c7c fix msvc C2398 error (#19500) Hu Shiwen 2020-11-12 12:37:00 +08:00
  • b9e8efac6a Update CONTRIBUTORS.md (#19518) Leonard Lausen 2020-11-12 01:47:31 +00:00
  • b3a6317391 Fix HLS2RGBConvert (#19516) Leonard Lausen 2020-11-11 22:25:03 +00:00
  • a52b6390c4 Optimize graph execution in presence of dynamic shape op (#18690) Manu Seth 2020-11-11 10:02:00 -08:00
  • 1120d5914b Fix markdown syntax for link to community guide (#19513) Yuan Tang 2020-11-11 09:08:35 -05:00
  • 351953c35c backport #19037 (#19514) waytrue17 2020-11-10 19:12:06 -08:00
  • ced029a299 Mark test_npx_batch_norm as flaky (#19508) Leonard Lausen 2020-11-10 21:48:04 +00:00
  • 495a1978ac Mark test_gluon_bernoulli_v1 as flaky (#19510) Leonard Lausen 2020-11-10 21:46:41 +00:00
  • a69e2b43ae Enable coverage reports for PR diffs (#19505) Leonard Lausen 2020-11-10 20:29:15 +00:00
  • f79e5bd86e Update CONTRIBUTORS.md (#19512) Sheng Zha 2020-11-10 15:08:59 -05:00
  • 74e5392253 Unique names for nodes in imperative mode (#19509) bgawrych 2020-11-10 20:41:10 +01:00