[PyTorch Debug] More advanced stats for Quantized Tensors #1897
Conversation
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
for more information, see https://pre-commit.ci
Signed-off-by: Pawel Gadzinski <pgadzinski@nvidia.com>
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
for more information, see https://pre-commit.ci
pggPL
force-pushed
the
debug_log_more_fp8_stats
branch
from
30688dd to
b78a504
Compare
for more information, see https://pre-commit.ci
timmoon10
reviewed
Jul 9, 2025
| @@ -231,6 +233,9 @@ def inspect_tensor( | |||
| tp_group: torch.distributed.ProcessGroup, | |||
| ) -> None: | |||
| """ | |||
| This is legacy call, we advise to use *inspect_tensor_all* and *inspect_tensor_all_enabled* instead. | |||
There was a problem hiding this comment.
If we don't want users to call inspect_tensor and other legacy functions, we should raise a deprecation warning.
Also, it seems like a shame to so quickly abandon these function names and replace them with clunky _all variants. Is there a way to maintain backward compatibility? Maybe have kwargs like high_precision_tensors and quantized_tensors, and have quantized_tensors=False by default.
| It allows to inspect both quantized and high precision tensors. | ||
| The feature LogFp8TensorStats uses this call to collect FP8 statistics after the quantization. | ||
|
|
||
| If tensor and the transpose are quantized differently |
There was a problem hiding this comment.
Accidentally deleted some of the docstring?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Description
This PR adds more statistics for Quantized Tensors that can be used to debug FP8 convergence issues.
It also adds inspect_tensor_all api call, which was necessary to support this features.
Type of change
Documentation change (change only to the documentation, either a fix or a new content)
Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to not work as expected)
Infra/Build change
Code refactoring
I have read and followed the contributing guidelines
The functionality is complete
I have commented my code, particularly in hard-to-understand areas
I have made corresponding changes to the documentation
My changes generate no new warnings
I have added tests that prove my fix is effective or that my feature works
New and existing unit tests pass locally with my changes