[AMP] Disallow fp16 conversion for arange op #8644
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Right now running the fp16 conversion pass on arange op results in a type error fp16 vs fp32, since we need to update
dtypeattributes of arangetvm/include/tvm/relay/attrs/transform.h
Line 197 in 1fac10b
We could fix the type error that way and support fp16 arguments to arange, but that would cause a problem when
endargument is a large number which happens to be fp32. So we should disallow fp16 conversion for arange. Since the inputs to arange are scalars, cast overhead is negligible.@AndrewZhaoLuo