Added support for TElinear ops #632
Open
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
iCo-authored-by: Asma Kuriparambil Thekkumpate <[email protected]> Co-authored-by: Kinjal Patel <[email protected]> Signed-off-by: Kinjal Patel <[email protected]>
Signed-off-by: Kinjal Patel <[email protected]>
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #632 +/- ##
==========================================
- Coverage 74.64% 74.55% -0.09%
==========================================
Files 183 183
Lines 18542 18432 -110
==========================================
- Hits 13840 13742 -98
+ Misses 4702 4690 -12 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Signed-off-by: Kinjal Patel <[email protected]>
realAsma
reviewed
Dec 3, 2025
realAsma
approved these changes
Dec 3, 2025
Signed-off-by: Kinjal Patel <[email protected]>
Signed-off-by: Kinjal Patel <[email protected]>
mxinO
reviewed
Dec 4, 2025
Contributor
mxinO
left a comment
mxinO
left a comment
There was a problem hiding this comment.
Looks great to me! Do we still need a special modelopt layer spec after this MR?
| ) | ||
|
|
||
|
|
||
| def test_convert_mcore_te_gpt_model(distributed_setup_size_1): |
Contributor
There was a problem hiding this comment.
Do we have basic tests for TE? ref_output == test_output for quantizers disabled, and ref_output != test_output for quantizers enabled?
| weight, weight_fp8, inputs = args[0], args[1], args[2] | ||
| remaining_args = args[3:] | ||
| idx = 1 if func_name == "_forward" else 0 | ||
| weight, weight_fp8, inputs = args[idx], args[idx + 1], args[idx + 2] |
Contributor
There was a problem hiding this comment.
We should raise an error somewhere if TE fp8 training is enabled when using modelopt.
Signed-off-by: Kinjal Patel <[email protected]>
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.
iCo-authored-by: Asma Kuriparambil Thekkumpate [email protected]
What does this PR do?
Type of change: New Feature
Overview:
This MR adds support for quantizing TE Ops in megatron, specifically TERowParallelLinear, TEColumnParallelLinear and TELayerNormColumnParallelLinear.
Usage
It can be used by enabling TE spec in megatron
Testing
Added unit tests for testing functionality
test_homogeneous_sharded_state_dict_te_spectest_convert_mcore_te_gpt_modeltest_quantize_forward_backwardBefore your PR is "Ready for review"
Additional Information