This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 132daf6c95 [Disco] Fix double free of nccl communicator (#17275)
add 4a37f64167 [KVCache] Increase coalesce threshold (#17280)
add 517c420d7b [TOPI][ADRENO] Add Group Conv2d texture schedule (#17274)
add 6bcec1d6c3 [CI] Resolve CI compilation failures on MacOSX (#17271)
add 6f4ac2312b [Relay][Pytorch] Add support for `aten::tile` (#17277)
add 1ca9833db2 [IR] Handle NaN in StructuralEqual and StructuralHash
(#17249)
add 7bea15f162 [WINDOWS] Compiler options for non x86 targets (#17260)
No new revisions were added by this update.
Summary of changes:
.github/actions/setup/action.yml | 4 +
conda/build_win.bat | 4 +-
conda/recipe/bld.bat | 2 +-
conda/recipe/install_libtvm.bat | 8 +-
conda/recipe/install_tvm_python.bat | 4 +-
include/tvm/node/structural_equal.h | 21 +++-
include/tvm/node/structural_hash.h | 13 +-
python/tvm/contrib/cc.py | 5 +-
python/tvm/dlight/gpu/gemv.py | 15 ++-
python/tvm/relay/frontend/pytorch.py | 11 ++
python/tvm/relay/op/strategy/adreno.py | 31 ++++-
python/tvm/topi/adreno/__init__.py | 1 +
.../{conv2d_nchw.py => group_conv2d_nchw.py} | 73 ++++++-----
src/runtime/relax_vm/paged_kv_cache.cc | 4 +-
tests/python/frontend/pytorch/test_forward.py | 24 ++++
...exture.py => test_group_conv2d_nchw_texture.py} | 133 +++++++--------------
.../tir-base/test_tir_structural_equal_hash.py | 43 +++++++
17 files changed, 261 insertions(+), 135 deletions(-)
copy python/tvm/topi/adreno/{conv2d_nchw.py => group_conv2d_nchw.py} (86%)
copy tests/python/relay/opencl_texture/{test_depthwise_conv2d_nchw_texture.py
=> test_group_conv2d_nchw_texture.py} (62%)
