This is an automated email from the ASF dual-hosted git repository.
github-actions[bot] pushed a change to branch nightly
in repository https://gitbox.apache.org/repos/asf/tvm.git
from 6b27d19497 [REFACTOR][IR] Remove dead AttrFunctor template (#19528)
add e370fc7374 [Relax][ONNX] Normalize negative indices before the take
call for `Gather` operator (#19525)
add 2c76c7955a [Relax][Frontend] Add TFLite Frontend Support for
CONV_3D_TRANSPOSE (#19530)
add c309e4ea5f [TIR] Add cooperative_tensor builtins and
metal.cooperative_tensor storage scope (#19423)
add c0406a54bc [Relax][Frontend][TFLite] Add initial StableHLO builtin
operator support (#19536)
add dfc9fc03b1 [Contrib] Fix CUDA contrib build after FFI/header cleanups
(#19539)
add 378c4f3043 [BugFix][Relax]: handle ONNX ScatterElements reduction
(#19527)
No new revisions were added by this update.
Summary of changes:
include/tvm/tirx/builtin.h | 45 +
python/tvm/relax/frontend/onnx/onnx_frontend.py | 59 +-
.../tvm/relax/frontend/tflite/tflite_frontend.py | 641 ++++++++++
python/tvm/tirx/op.py | 104 ++
python/tvm/tirx/script/builder/ir.py | 8 +
.../cutlass/fp16_group_gemm_runner_sm100.cuh | 2 +
.../cutlass/fp16_group_gemm_runner_sm90.cuh | 2 +
.../contrib/cutlass/fp8_groupwise_scaled_gemm.cuh | 1 +
src/runtime/contrib/cutlass/gemm_runner.cuh | 2 +
src/runtime/contrib/nvshmem/init.cc | 1 +
src/runtime/contrib/thrust/thrust.cu | 1 +
src/runtime/contrib/vllm/attention_kernels.cu | 49 +-
src/runtime/contrib/vllm/cache_kernels.cu | 4 +-
src/runtime/thread_storage_scope.h | 7 +
src/tirx/op/builtin.cc | 12 +
tests/python/relax/test_frontend_onnx.py | 162 +++
tests/python/relax/test_frontend_tflite.py | 1236 ++++++++++++++++++++
17 files changed, 2313 insertions(+), 23 deletions(-)
