- 14 Aug, 2017 5 commits
-
-
Nicolas Vasilache committed
-
* add dilation operators * fix pylint * dilate testcases success * n-D tensor dilation * support arbitrary dimension
Yuwei HU committed -
* [DOC] Include TOPI in doxygen * update
Tianqi Chen committed -
* __init__ updated * pull request updated * build_module added * typo fixed * another typo fixed * conv2d gpu scheduler for two layouts moved to tvm * changes made according to CR * conv2d_nchw formating updated, conv2d_hwcn tests updated * lint error fixed * element wise operator schedule fusing fixed for conv2d * conv2d_nchw topi test added, all resnet workloads now pass * conv compute lint error fixed * fixed python 3 compatibility problem * conv2d tensor input support added, test typo fixed, ir_pass.Simplify changed to util.get_const_int
Leyuan Wang committed -
* [TOPI] Move topi.nn.util to topi.util * update the path
Tianqi Chen committed
-
- 13 Aug, 2017 3 commits
-
-
* [WIP] C++ topi contributions Summary: This diff implements C++ topi contributions for: - relu with parametrix threshold - pad with generic padBefore / padAfter specification - matmult with transposes - conv2d_nchw, conv2d_hwcn with runtime constant padding and strides - depthwise_conv2d_nchw with runtime constant padding and strides - group_conv2d_ngchw with runtime constant padding and strides - broadcast_to a broadcastable shape - broadcast_bop where bop is an usual binary op (+ - * / %) Convolution padding is implemented using the pad operation. To avoid extra memory consumption, it is generally recommended to inline the padding with the autoinliner. Unfortunately in its current form the elemwise checks are too restrictive to allow inlining. So this diff also proposes an extension to LHS injective (i.e. no reduction axis in the current IR design) Test Plan: Tested in C++ testsuite in a separate repository, I am looking for suggestions to quickly spin up some tests for tvm. Reviewers: tqchen Subscribers: Tasks: Tags: Blame Revision: * Review + Lint + GSG C++
Nicolas Vasilache committed -
Tianqi Chen committed
-
Tianqi Chen committed
-
- 12 Aug, 2017 1 commit
-
-
Tianqi Chen committed
-
- 11 Aug, 2017 2 commits
-
-
Yuwei HU committed
-
* [PASS][FIX] Fix LiftAttrScope with if * [PASS] Fix on proc sync * fix
Tianqi Chen committed
-
- 10 Aug, 2017 3 commits
-
-
Tianqi Chen committed
-
Yizhi Liu committed
-
* [TEST] Upgrade gpu docker to cudnn7 * fx
Tianqi Chen committed
-
- 09 Aug, 2017 2 commits
-
-
Tianqi Chen committed
-
ziheng committed
-
- 08 Aug, 2017 3 commits
-
-
Tianqi Chen committed
-
* [tvm4j] RPC Server * [tvm4j] fix recursively function calling; connect to proxy server; osx rename .so to .dylib * [tvm4j] test case for proxy connection; thread pool for serving
Yizhi Liu committed -
* [RUNTIME][PASS] Allow declare vector type array * fix bcast * [BUFFER] Enable vload/store function in buffer * ok
Tianqi Chen committed
-
- 07 Aug, 2017 1 commit
-
-
* [NNPACK] Add nnpack.convolution * Add instrinsic * Fix lint
ziheng committed
-
- 06 Aug, 2017 1 commit
-
-
* [EXECUTOR] Split graph_executor to header file and (runtime) source file * Fix
ziheng committed
-
- 05 Aug, 2017 1 commit
-
-
Tianqi Chen committed
-
- 04 Aug, 2017 2 commits
-
-
Tianqi Chen committed
-
* [PASS] Enhance LayoutTransform pass * Fix * Fix Compilation * Refactor * Refactor * doc * fix * add file
ziheng committed
-
- 03 Aug, 2017 2 commits
-
-
* [PASS] Refactor thread storage sync to a common visitor * Fix the sync scope check behavior
Tianqi Chen committed -
Tianqi Chen committed
-
- 01 Aug, 2017 6 commits
-
-
* Typofix. Signed-off-by: Edward Z. Yang <ezyang@fb.com> * Probe for nvrtc in lib directory as well. Signed-off-by: Edward Z. Yang <ezyang@fb.com> * Conda build recipe for TVM. Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Edward Z. Yang committed -
ziheng committed
-
Tianqi Chen committed
-
Yizhi Liu committed
-
* [PASS] Improve graph fusion * Change fusion center to segment head * Use 'master' to identity the schedule node * Make things compact * Fix
ziheng committed -
* [SCHEDULE] Fix fuse node order * Make fuse order consistent with split
Tianqi Chen committed
-
- 31 Jul, 2017 3 commits
-
-
* Install rules and dynamic library loading fixes. A batch of fixes: - Added 'install' rule to cmake and make, which installs runtime headers and library (libtvm_runtime). - Added 'installdev' rule to make, which also installs the compiler infrastructure headers and library (libtvm) - Added 'INSTALL_DEV' option to cmake, for toggling installation of compiler infrastructure headers and library - cmake no longer builds into lib/ directory; instead all build products go in your build directory - New algorithm for dynamic library loading, as described in #281. Signed-off-by: Edward Z. Yang <ezyang@fb.com> * Emit dylib on OS X Signed-off-by: Edward Z. Yang <ezyang@fb.com> * Lint fixes. Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Edward Z. Yang committed -
Mercy committed
-
William Moses committed
-
- 29 Jul, 2017 1 commit
-
-
* __init__ updated * pull request updated * build_module added * typo fixed * another typo fixed
Leyuan Wang committed
-
- 28 Jul, 2017 1 commit
-
-
* [GRAPH] Add GraphHandle * Move to apps/graph_executor
ziheng committed
-
- 27 Jul, 2017 3 commits
-
-
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Edward Z. Yang committed -
Tianqi Chen committed
-
* image add! * image path move to web-data
Jian Weng committed
-