Fix import of TE Recipe - #2635
Merged
Merged
Conversation
Signed-off-by: Kirthi Shankar Sivamani <ksivamani@nvidia.com>
ksivaman
requested review from
KaelanDt,
lantiga,
mruberry and
t-vi
as code owners
October 13, 2025 18:03
Collaborator
|
Adding @riccardofelluga as reviewer. Please take a look |
t-vi
approved these changes
Oct 13, 2025
13 tasks
Comment on lines
+33
to
-36
| from transformer_engine.common.recipe import Recipe | ||
| from transformer_engine.pytorch.fp8 import ( | ||
| _amax_and_scale_update, | ||
| get_fp8_max, | ||
| Recipe, |
Collaborator
There was a problem hiding this comment.
Thanks for the update! LGTM
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
4 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Before submitting
What does this PR do?
Thunder uses internal TE API which is being refactored in NVIDIA/TransformerEngine#2259. The TE PR ensures full backwards compatibility, however, the
Recipeclass was not defined infp8.pyto begin with and this only works because it is imported indirectly in it. This PR imports it directly in order to not break integration immediately.Future work
Use of internal TE API should be minimized, and imports from
fp8.pyshould be changed as it will be eventually removed (currently renamed toquantization.pystarting TE v2.9).