Add bfloat16 support to JaggedTensor reduce operators#501
Merged
Conversation
Add c10::kBFloat16 to the AT_DISPATCH_V2 type list in JaggedReduce.cu, enabling jsum/jmin/jmax to accept bfloat16 tensors. JaggedSort already supported bfloat16; JaggedReduce was the only jagged kernel missing it. Expand test_jagged_tensor.py to cover bfloat16 on both CUDA and CPU, with appropriately wider tolerances and smaller dataset sizes to account for bfloat16's 7-bit mantissa. Skip tests that depend on grid-building ops (GridBatch.from_points) which do not yet support bfloat16. Signed-off-by: Jonathan Swartz <jonathan@jswartz.info>
swahtz
added
the
core library
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add bfloat16 support to JaggedTensor reduce operators: add c10::kBFloat16 to the AT_DISPATCH_V2 type list in JaggedReduce.cu, enabling jsum/jmin/jmax to accept bfloat16 tensors. JaggedSort already supported bfloat16; JaggedReduce was the only jagged kernel missing it.
Expand test_jagged_tensor.py to cover bfloat16 on both CUDA and CPU, with appropriately wider tolerances and smaller dataset sizes to account for bfloat16's 7-bit mantissa. Skip tests that depend on grid-building ops (GridBatch.from_points) which do not yet support bfloat16.
fixes #395