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 7b28787819 [FFI] Update torch stream getter to use native torch c api
(#18266)
add 536091010b [FFI] Support Opaque PyObject (#18270)
add 86b391a4b6 [FFI] Support inline module (#18271)
No new revisions were added by this update.
Summary of changes:
ffi/examples/inline_module/main.py | 86 +++++
ffi/include/tvm/ffi/c_api.h | 199 ++++++-----
ffi/python/tvm_ffi/convert.py | 6 +-
.../python/tvm_ffi/cpp/__init__.py | 6 +-
ffi/python/tvm_ffi/cpp/load_inline.py | 382 +++++++++++++++++++++
ffi/python/tvm_ffi/cython/base.pxi | 7 +
ffi/python/tvm_ffi/cython/function.pxi | 30 +-
ffi/python/tvm_ffi/cython/object.pxi | 17 +
.../python/tvm_ffi/utils}/__init__.py | 3 +-
ffi/python/tvm_ffi/utils/lockfile.py | 113 ++++++
ffi/src/ffi/object.cc | 40 +++
ffi/tests/cpp/test_object.cc | 25 ++
ffi/tests/python/test_container.py | 22 ++
ffi/tests/python/test_function.py | 25 ++
ffi/tests/python/test_load_inline.py | 161 +++++++++
ffi/tests/python/test_object.py | 21 ++
16 files changed, 1044 insertions(+), 99 deletions(-)
create mode 100644 ffi/examples/inline_module/main.py
copy conda/recipe/conda_build_config.yaml =>
ffi/python/tvm_ffi/cpp/__init__.py (95%)
create mode 100644 ffi/python/tvm_ffi/cpp/load_inline.py
copy {python/tvm/relax/backend/contrib =>
ffi/python/tvm_ffi/utils}/__init__.py (96%)
create mode 100644 ffi/python/tvm_ffi/utils/lockfile.py
create mode 100644 ffi/tests/python/test_load_inline.py
