saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-26 19:52:38 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
onnxruntime/orttraining/orttraining/python
History
Vincent Wang 9f68a27c7a
[ORTModule] Handle Cast on Constant Number on Triton Code-gen (#19321) 
When using scaled_dot_product_attention on float16 type, the exported
graph has Sqrt(float16(constant)), which cannot be ConstantFold in ORT
because Sqrt CPU kernel doesn't support float16. This causes Triton
code-gen generates code like:

result = 128.0.to(tl.float32)

This code cannot be compiled because .to() cannot be applied to
constant.

This PR is to handle such case that constant number will not do the
Cast.
2024-01-30 17:04:01 +08:00
..
training [ORTModule] Handle Cast on Constant Number on Triton Code-gen (#19321) 2024-01-30 17:04:01 +08:00
orttraining_pybind_common.h
orttraining_pybind_state.cc Improve perf for stage3 training (#18099) 2023-12-15 13:32:19 +08:00
orttraining_python_module.cc Fix warning C4003 in ORT python binding code (#18612) 2023-11-30 08:07:47 -08:00
orttraining_python_module_eager.h Run clang-format in CI (#15524) 2023-04-18 09:26:58 -07:00
pt_patch.py Adopt linrtunner as the linting tool - take 2 (#15085) 2023-03-24 15:29:03 -07:00