- 29 May, 2018 40 commits
-
-
Tianqi Chen committed
-
* initial NHWC layout support * remove layout param from softmax * more nhwc support * fix typo * add nhwc layout test * fix lint * update tvm * update for c++ topi * fix lint * update tvm
masahi committed -
Siva committed
-
The below compilation warning is fixed. In file included from /mnt/D_DRIVE/work/nnvm_8_feb/dmlc-core/include/dmlc/any.h:16:0, from include/nnvm/./base.h:11, from include/nnvm/graph.h:15, from src/compiler/fold_scale_axis.cc:6: src/compiler/fold_scale_axis.cc: In function ‘nnvm::Graph nnvm::compiler::FoldScaleAxis(nnvm::Graph)’: src/compiler/fold_scale_axis.cc:155:39: warning: enum constant in boolean context [-Wint-in-bool-context] CHECK(kind == kPassTroughFirst || kMulConsumer); ^ /mnt/D_DRIVE/work/nnvm_8_feb/dmlc-core/include/dmlc/./logging.h:110:9: note: in definition of macro ‘CHECK’ if (!(x))
Siju Samuel committed -
Yao Wang committed
-
* Updated TVM version. Implemented fix for nnvm_compiler crash on exit on windows. Changed TOPI ops from using python to using C++ where applicable. * Fix lint * Fix lint * Fix macro * Fix reshape * Update TVM to fix test fails
alex-weaver committed -
The below compilation warning issue has been fixed. src/top/tensor/matrix_op.cc:37:21: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (int i = 0; i < lshape.ndim() - 1; i++) oshape[i] = lshape[i]; ^ src/top/tensor/matrix_op.cc:38:21: warning: comparison between signed and unsigned integer expressions [-Wsign-compare] for (int i = 1; i < rshape.ndim(); i++) oshape[i + lshape.ndim() - 1] = rshap
Siju Samuel committed -
fix similar problem like https://github.com/apache/incubator-mxnet/pull/4589/files .
Liangfu Chen committed -
* add CacheItem2Schedule extension * fix lint * move function position * make cache item visible to frontend
Lianmin Zheng committed -
Siva committed
-
* Add conv2d max_pool backward op * Added tests * Fix testing * Address comments * Change dot to matmul * Address comments * Break down indicator function * Make greater, less numpy compatible
yuruofeifei committed -
* make shape inference of BatchNorm layout neutral * refactor to use the axis variable to do BatchNorm shape inference * refactor to use the axis variable to do BatchNorm shape inference * add unittest to the axis param for batch norm shape inference
guopinglong committed -
* add nnvm upsampling symbol * add upsampling mxnet frontend * add doc for upsampling op * cleanup upsampling test * minor fix * use schedule_injective for upsampling * upgrade tvm
masahi committed -
f9ed337552a26cc55c7c0fb22cd54839ee13f19c added a call to std::log without including cmath
Daniel M. Weeks committed -
ziheng committed
-
* Add creating gradient symbol * Fix lint * Address comments * Fix typo * Address comment
Yao Wang committed -
* Update docs * Add backward operator to enable backward graph * Fix testing * Refactor top level1 test code * Fix format * Test * Added zeros ones op * Register fill_like operator * Fix unit test
yuruofeifei committed -
* keras frontend tutorial * fix
Yuwei Hu committed -
* vgg16 success * remove six.PY2, use sys.version_info; convert_activation() accepts activation type name(str, e.g. 'relu') as input; * add convert_merge * fix convert_batchnorm; improve tests * fix lint * add numpy-style pad operator * deal with asymmetry padding * resnet50 success * fix pool_convert; xception passes test * update tvm * fix bias error; all tests pass * use > >, not >>
Yuwei Hu committed -
abergeron committed
-
* [SYMBOL] Add __iter__ and GetChildren for symbol * [SYMBOL] Fix lint
ziheng committed -
* Symbol __getitem__ using list_outputs() is too expensive, when it only cares about the output count in most cases * Add cython cmake * GetNumOutputs() and __len__ changes per PR comments * set commit for tvm
Chris Olivier committed -
* [CMPL] Add Support for Other Data Types * [CMPL] Add test * [CMPL] Fix
ziheng committed -
Lianmin Zheng committed
-
* update tvm * [TOP] fix weight layout in conv2d_transpose
Yuwei Hu committed -
* [TOP] add conv2d_transpose * update tvm * fix pylint
Yuwei Hu committed -
ziheng committed
-
Eric Junyuan Xie committed
-
Tianqi Chen committed
-
Tianqi Chen committed
-
Tianqi Chen committed
-
Tianqi Chen committed
-
* [PASS] FoldScaleAxis * Move FoldAxis to O3 * Set unroll to 0 when ready
Tianqi Chen committed -
* add transform * fix * update doc * Update tvm
Xingjian Shi committed -
Tianqi Chen committed
-
Tianqi Chen committed
-
tqchen committed
-
Tianqi Chen committed
-
Tianqi Chen committed
-
* [TUTORIAL] Move mobilenet to tutorial, fix precompute_prune * Some language improvements
Tianqi Chen committed
-