fix lm eval test reproducibility issues #1260
Merged
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
|
👋 Hi! Thank you for contributing to llm-compressor. Please add the ready label when the PR is ready for review. Note: This is required to complete the testing suite, please only add the label once the PR is code complete and local testing has been performed. |
dsikka
reviewed
Mar 17, 2025
|
@brian-dellabetta The down_proj is the most likely to fail hessian inversion, since it is the weight with the largest input size. This is likely a normal hessian inevitability issue, which can be fixed by shuffling the dataset/ using an image dataset |
|
Hitting several issues around reproducibility and slowness of lm-eval (takes about a minute for each of the 30 samples to run). Will continue this next week after resolving more urgent tasks |
dsikka
pushed a commit
that referenced
this pull request
Mar 21, 2025
SUMMARY: multi-modal lm-eval tests are failing due to a non-reproducibility issue that still needs to be resolved. In the meantime, moving those tests to a skipped folder until resolution. Resolution can be tracked in #1260 TEST PLAN: no new source code Signed-off-by: Brian Dellabetta <[email protected]>
brian-dellabetta
force-pushed
the
bdellabe/lmeval-multimodal-test-fixes
branch
from
07e5877 to
596f562
Compare
dbarbuzzi
reviewed
May 2, 2025
brian-dellabetta
force-pushed
the
bdellabe/lmeval-multimodal-test-fixes
branch
from
019c6bb to
6884ec6
Compare
brian-dellabetta
changed the title
kylesayrs
requested changes
May 5, 2025
tests/e2e/vLLM/recipes/actorder/recipe_w4a16_actorder_weight_dampfrac1e-1.yaml
Outdated
Show resolved
Hide resolved
kylesayrs
reviewed
May 5, 2025
kylesayrs
previously approved these changes
May 5, 2025
brian-dellabetta
commented
May 6, 2025
dsikka
approved these changes
May 6, 2025
There was a problem hiding this comment.
LGTM. Can you verify with Domenic that there will not be any conflicts with the existing tests that we're now tracking here? https://fantastic-adventure-plymqoj.pages.github.io/timings/lmeval/
I think it should be fine since we use the config name but just in case
Confirmed with domenic here, I will ask rahul to approve and then merge this in |
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
Signed-off-by: Brian Dellabetta <[email protected]>
brian-dellabetta
force-pushed
the
bdellabe/lmeval-multimodal-test-fixes
branch
from
56c12fa to
823fdd5
Compare
shanjiaz
pushed a commit
that referenced
this pull request
May 6, 2025
SUMMARY: lm-eval multimodal tests were failing to reproduce across different versions of compressed tensors. After upgrading the models from 2B to 7B, the tests appear to be reproducing across compressed tensors 0.9.1, 0.9.2 and nightly. I ran extensively for the fp8 config across different versions of CT, and it always returned the same result. I also removed the random seed from the configs. after running several of each of the 3 configs, i did not see any change in result. this may cause errors during ci/cd testing but I'd like to see if it does, i feel that is a better e2e test anyway. Tests take roughly 1hr30m - 1h45m to run. TEST PLAN: no new src code, just fixing tests --------- Signed-off-by: Brian Dellabetta <[email protected]> Signed-off-by: shanjiaz <[email protected]>
shanjiaz
pushed a commit
that referenced
this pull request
May 6, 2025
SUMMARY: lm-eval multimodal tests were failing to reproduce across different versions of compressed tensors. After upgrading the models from 2B to 7B, the tests appear to be reproducing across compressed tensors 0.9.1, 0.9.2 and nightly. I ran extensively for the fp8 config across different versions of CT, and it always returned the same result. I also removed the random seed from the configs. after running several of each of the 3 configs, i did not see any change in result. this may cause errors during ci/cd testing but I'd like to see if it does, i feel that is a better e2e test anyway. Tests take roughly 1hr30m - 1h45m to run. TEST PLAN: no new src code, just fixing tests --------- Signed-off-by: Brian Dellabetta <[email protected]> Signed-off-by: shanjiaz <[email protected]>
shanjiaz
pushed a commit
that referenced
this pull request
May 7, 2025
SUMMARY: lm-eval multimodal tests were failing to reproduce across different versions of compressed tensors. After upgrading the models from 2B to 7B, the tests appear to be reproducing across compressed tensors 0.9.1, 0.9.2 and nightly. I ran extensively for the fp8 config across different versions of CT, and it always returned the same result. I also removed the random seed from the configs. after running several of each of the 3 configs, i did not see any change in result. this may cause errors during ci/cd testing but I'd like to see if it does, i feel that is a better e2e test anyway. Tests take roughly 1hr30m - 1h45m to run. TEST PLAN: no new src code, just fixing tests --------- Signed-off-by: Brian Dellabetta <[email protected]> Signed-off-by: shanjiaz <[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.
SUMMARY:
lm-eval multimodal tests were failing to reproduce across different versions of compressed tensors. After upgrading the models from 2B to 7B, the tests appear to be reproducing across compressed tensors 0.9.1, 0.9.2 and nightly. I ran extensively for the fp8 config across different versions of CT, and it always returned the same result.
I also removed the random seed from the configs. after running several of each of the 3 configs, i did not see any change in result. this may cause errors during ci/cd testing but I'd like to see if it does, i feel that is a better e2e test anyway.
Tests take roughly 1hr30m - 1h45m to run.
TEST PLAN:
no new src code, just fixing tests