Commit Graph

  • da95add57b Fix vector access out of bound in MKLDNNConvolutionBackward (#17997) Tao Lv 2020-04-09 02:28:33 +08:00
  • 13841ddd2d [mkldnn] optimize for mkldnn batchnorm backward (#17902) rongzha1 2020-04-08 12:57:29 +08:00
  • b7f7525cba dnnl v1.2.2 (#17991) Tao Lv 2020-04-08 09:59:07 +08:00
  • 16ddc6d9c6 Custom Operator Random Number Generator Support (#17762) Ziyi Mu 2020-04-07 17:39:14 -07:00
  • f906a02b77 ffi_atleast_1/2/3d (#17897) Minghao Liu 2020-04-08 08:02:26 +08:00
  • a960f5a6eb ffi_array_split, v/h/dsplit (#17873) Minghao Liu 2020-04-08 08:02:00 +08:00
  • 892f9823e1 * impl - linalg.lstsq for cpu (#17950) dw_sjtu 2020-04-08 08:00:23 +08:00
  • 79c576b815 [ONNX export] Fixing spatial export for batchnorm (#17711) Vinitra Swamy 2020-04-07 13:42:58 -07:00
  • 002d4f1d91 * impl - FFi for linalg op (#17795) dw_sjtu 2020-04-07 14:39:53 +08:00
  • c3c76a8ba3 Optimize AddTakeGrad Tensor Sum (#17906) YixinBao 2020-04-07 08:48:31 +08:00
  • 5adcbf8384 GPU gemms true fp16 (#17466) MoisesHer 2020-04-06 17:15:08 -07:00
  • c244f9f129 [MXNET-#16795] Byteps-KVStore: Intergrate Byteps into mxnet as new type of kvstore backend (#17555) Chaokun Chang 2020-04-07 08:14:11 +08:00
  • ff234dbe1a Skip test_gluon_data.py on OSX (#17969) Leonard Lausen 2020-04-06 23:05:01 +00:00
  • 03b8146e62 Skip test_kvstore_gpu.test_rsp_push_pull (#17983) Leonard Lausen 2020-04-06 19:51:50 +00:00
  • 84b0ddd022 Add USE_DIST_KVSTORE=ON to GPU build (#17911) Lin Yuan 2020-04-06 12:49:41 -07:00
  • 1b107a0a0f Remove redundant condition in np_matrix_op.cc (#17933) ga 2020-04-06 14:54:53 -04:00
  • 4c0f763893 Fix a typo (#17979) Gaspar Vardanyan 2020-04-06 20:32:37 +04:00
  • 2fff11d423 [OpPerf] Fix axis_shape and function mismatch for LTS (#17894) Chaitanya Prakash Bapat 2020-04-05 13:12:19 -07:00
  • 21fc103d99 [Website 2.0] Nightly Build for v1.x (#17956) Connor Goggins 2020-04-04 16:15:11 -07:00
  • b6edefbf3e CI: Re-enable code coverage for CPU builds (#17889) Leonard Lausen 2020-04-04 01:51:56 +00:00
  • 792011e73d Omit kNullOp req when comparing changed NDArrays in static_shape=True (#17966) Przemyslaw Tredak 2020-04-03 15:35:52 -07:00
  • 66ee118a45 Fix Windows GPU CI (#17962) Leonard Lausen 2020-04-03 20:20:47 +00:00
  • 08528c5cb9 [numpy] add op median (#17084) Yiyan66 2020-04-02 12:03:36 +08:00
  • 56e79853ad * impl - FFI for np einsum (#17869) dw_sjtu 2020-03-26 01:31:18 +08:00
  • d4052fde4a [Numpy] FFI for diag/diagonal/diag_indices_from (#17789) Minghao Liu 2020-03-24 15:26:32 +08:00
  • 9a355ebc1d [Numpy] Kron operator (#17323) hanke580 2020-03-24 03:41:58 +08:00
  • 3840786a25 cmake: Set DMLC_LOG_FATAL_THROW only for building mxnet and not for tvm (#17878) Leonard Lausen 2020-03-23 11:25:26 -07:00
  • 83b51703ed Add simplified HybridBlock.forward without F (#17530) Leonard Lausen 2020-03-23 11:21:23 -07:00
  • 8e3951876b Use FP32 copy of weights for norm (multitensor LAMB optimizer) (#17700) MoisesHer 2020-03-23 09:55:24 -07:00
  • b1338999c1 Use multi-tensor sumSQ in clip_global_norm (#17652) MoisesHer 2020-03-23 09:51:41 -07:00
  • 2f358fdc31 [Numpy] Add op fmax, fmin, fmod (#17567) hanke580 2020-03-23 15:36:39 +08:00
  • f01dc80f03 Adding sparse support to MXTensor for custom operators (#17569) guanxinq 2020-03-22 03:50:55 -07:00
  • 3667e9a205 Update 3rdparty/mkldnn to v1.2.2 (#17313) Tao Lv 2020-03-21 00:12:19 +08:00
  • f7c43234d0 Dynamic subgraph compile support (#17623) Sam Skalicky 2020-03-19 12:38:59 -07:00
  • 4fa4e65700 Refactor cpp-package CMakeLists.txt & add missing inference/imagenet_inference (#17835) Leonard Lausen 2020-03-18 19:39:31 -07:00
  • 04167dfffd staticbuild: Fix potential user-assisted execution of arbitrary code (#17860) Leonard Lausen 2020-03-18 19:38:24 -07:00
  • ea2320a897 * FFI for np.argmax and np.argmin (#17843) dw_sjtu 2020-03-19 02:37:29 +08:00
  • 07d0b73838 ffi for roll/rot90 (#17861) Minghao Liu 2020-03-19 02:10:41 +08:00
  • 41d534be3f Skip test_multi_worker_dataloader_release_pool on OS X (#17797) Hao Jin 2020-03-18 10:14:47 -07:00
  • 2fae7e43c2 add ffi for full_like, binary (#17811) alicia 2020-03-18 15:14:21 +08:00
  • dfb1b882d0 HybridBlock.export() to return created filenames (#17758) Alex Thewsey 2020-03-18 12:44:17 +08:00
  • 599654402a Fix SoftReLU fused operator numerical stability (#17849) RuRo 2020-03-18 07:42:53 +03:00
  • ab4f7f6a73 CI: Test clang10 cpu & gpu builds with -WError (#17830) Leonard Lausen 2020-03-17 21:36:50 -07:00
  • 1efb747ad5 [Numpy] FFI for split and svd (#17816) Haozheng Fan 2020-03-18 05:06:01 +08:00
  • a7ecb35738 ffi wrappers for polyval, ediff1d, nan_to_num (#17832) alicia 2020-03-18 05:05:34 +08:00
  • 796fa50a46 * impl - FFI for np_where_op (#17817) dw_sjtu 2020-03-18 00:36:36 +08:00
  • 1368a084b5 CI: switch to patched ccache 3 version (#17853) Leonard Lausen 2020-03-16 23:24:45 -07:00
  • 66b21b5bcd fixing batch_norm and layer_norm for large tensors (#17805) Rohit Kumar Srivastava 2020-03-16 16:02:05 -07:00
  • bd6b80edd0 CI: Disable AVX instructions in libzstd ccache dependency (#17842) Leonard Lausen 2020-03-16 10:28:47 -07:00
  • ea2dabaae5 [NumPy] add op random.laplace (#17316) AntiZpvoh 2020-03-16 12:52:53 +08:00
  • d495f228eb fix chisquare (#17825) Xi Wang 2020-03-15 13:38:01 +08:00
  • bbba4622e8 Update ccache on CI to fix cuda object caching (#17828) Leonard Lausen 2020-03-14 16:55:03 -07:00
  • c5989fa590 [CD] update pypi description, setup.py (#17681) Sheng Zha 2020-03-13 11:56:50 -07:00
  • 34010ea417 [CD] switch CD_RELEASE_JOB_NAME from global env var to job argument (#17775) Sheng Zha 2020-03-12 23:11:49 -07:00
  • 18c2a264bb [numpy] add magic methods for symbol bitwise ops (#17807) Yiyan66 2020-03-13 09:51:08 +08:00
  • dfbcf6f8a1 fix np.clip scalar input case (#17788) Xi Wang 2020-03-13 02:14:28 +08:00
  • bd6e917611 [numpy] add op random.f (#17586) Yiyan66 2020-03-12 16:47:53 +08:00
  • 713d962359 [Numpy] FFI: Bincount, Percentile/Quantile, All/Any (#17717) Minghao Liu 2020-03-11 06:02:00 +08:00
  • a5d55dea05 [OpPerf] Consolidate array manipulation related operators (#17487) Chaitanya Prakash Bapat 2020-03-10 14:01:49 -07:00
  • afb8742e6e Add ffi benchmark (#17780) Haozheng Fan 2020-03-10 13:37:18 +08:00
  • 4dddb08729 Implement storage tagging, the first half of the memory profiler (#17656) Bojian Zheng 2020-03-09 23:36:30 -04:00
  • 2cf8e0f7de ffi invocation: expand_dims, tril, diff, broadcast_to (#17738) alicia 2020-03-10 06:28:04 +08:00
  • b001006f8f add npx.broadcast_like (#17605) ZheyuYe 2020-03-10 05:13:55 +08:00
  • 0aa2c78cfb Fix OpPerf in Master (#17735) Chaitanya Prakash Bapat 2020-03-09 13:46:03 -07:00
  • 517e3a1a29 Fix mxnet-native and Docker CD pipelines (#17784) Leonard Lausen 2020-03-09 13:41:46 -07:00
  • 752461d172 refactor codes and add an option to skip/check weight's version to reduce overhead (#17707) ciyong 2020-03-09 10:06:10 +08:00
  • 1bd0a08419 [CI] Remove print stacktrace (due to security concerns) (#17597) Chaitanya Prakash Bapat 2020-03-08 05:11:41 -07:00
  • 27c74ade48 [Numpy] Fix symbolic basic indexing (#17770) Xingjian Shi 2020-03-06 14:17:33 -08:00
  • 3b83cd8036 Bump up additional scala 1.x branch to 1.7.0 (#17765) Zach Kimberg 2020-03-06 13:39:37 -08:00
  • 592466e96a allclose_op: Fix non-ascii characters in comments (#17746) Bairen Yi 2020-03-07 02:34:11 +08:00
  • e4c5f8c70e Expose activation and recurrent activation in GRUCell (#17745) Ehsan M. Kermani 2020-03-06 09:51:36 -08:00
  • cfb474ba74 add alias for np.__version__, np._NoValue and np.dtype (#17777) Hao Jin 2020-03-05 20:34:56 -08:00
  • 48e9e2c6a1 bump up master branch to 2.0 (#17739) Sheng Zha 2020-03-05 17:29:24 -08:00
  • d02115fca9 add logic for no batch size while getting data arrays from executors (#17772) Manu Seth 2020-03-06 06:59:06 +05:30
  • 7c61cdadeb Fix CD (#17776) Leonard Lausen 2020-03-05 11:20:34 -08:00
  • 66e4c27de6 cmake: improve CMAKE_CUDA_COMPILER guessing on unix (#17773) Leonard Lausen 2020-03-05 09:10:43 -08:00
  • 938b35b0d3 [Numpy] FFI for cumsum and add (#17747) Hao Jin 2020-03-04 22:06:46 -08:00
  • aba400846d [CI] Run python unittests on OS X with Github Actions (#17771) Leonard Lausen 2020-03-04 21:15:18 -08:00
  • 49f94d8601 Update get_started page for master cpu build (#17757) Leonard Lausen 2020-03-04 16:11:44 -08:00
  • 64d8e0f598 Fix setting CMAKE_GENERATOR "Ninja" in config.cmake (#17742) Leonard Lausen 2020-03-04 08:55:56 -08:00
  • 636897718f Fix MKL static link & default to static link on unix (#17751) Leonard Lausen 2020-03-04 08:55:01 -08:00
  • 5cffa74485 [Large Tensor] Fixed RNN op (#17632) Connor Goggins 2020-03-03 22:29:45 -08:00
  • 21b7fc028d Add 1.6.0 to the list of signed download links (#17753) Przemyslaw Tredak 2020-03-03 22:26:15 -08:00
  • af322dc802 [CI] Test gcc8 -WError build CI (#17752) Leonard Lausen 2020-03-03 21:01:59 -08:00
  • 2abd2255cf Update website, README and NEWS with 1.6.0 (#17658) Przemyslaw Tredak 2020-03-03 10:15:39 -08:00
  • 9b7132b7ac change error tolerance for bf16 bn (#17673) rongzha1 2020-03-03 14:03:01 +08:00
  • 9ed5a38651 [Large Tensor] Fix multi_lars op (#17675) Connor Goggins 2020-03-02 22:02:10 -08:00
  • 91d595a4bc bump up 1.x branch to 1.7.0 (#17741) Sheng Zha 2020-03-02 21:22:46 -08:00
  • 458906b9c6 bump up 1.x branch to 1.7.0 (#17741) Sheng Zha 2020-03-02 21:22:46 -08:00
  • e7e5b40251 add numpy_fallback decorator (#17657) Hao Jin 2020-03-02 21:01:17 -08:00
  • 4e2305f521 more support for boolean indexing and assign (#17672) alicia 2020-03-03 13:00:28 +08:00
  • c2e9a88568 Fix nightly tests pipeline (#17721) Leonard Lausen 2020-03-02 15:29:10 -08:00
  • 04ae389878 Update 3rdparty/ps-lite (#17732) Leonard Lausen 2020-03-02 14:27:53 -08:00
  • 04fe114d7b sum_fix (#17737) Minghao Liu 2020-03-03 06:00:33 +08:00
  • ac7797437e Support projection feature for LSTM on CPU (Only Inference) (#17702) Zixuan Wei 2020-03-02 09:31:11 +08:00
  • 10a12d59f6 Fix get_started pip instructions (#17725) Leonard Lausen 2020-02-29 13:09:12 -08:00
  • f70c7b7b1e [MXNET-#16167] Refactor Optimizer (#17400) Shuai Zheng 2020-02-29 09:25:43 -08:00
  • 88b3051f29 [MKLDNN] apply MKLDNNRun to quantized_act/transpose (#17689) Wuxun Zhang 2020-02-29 21:17:38 +08:00
  • 95c5189cc7 [Large Tensor] Implemented LT flag for OpPerf testing (#17449) Connor Goggins 2020-02-29 00:43:08 -08:00
  • 2527553e8c [Large Tensor] Fix cumsum op (#17677) Connor Goggins 2020-02-29 00:37:33 -08:00