Add AWQ/SmoothQuant evaluation wrapper test for vLLM benchmark #3000
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
Introduce a prototype AWQ/SmoothQuant benchmark within lm_eval (vLLM). Related Issue/PR: pytorch#2815 Test plan: A toy tokenizer model is introduced to validate `TransformerEvalWrapper`. Toy linear models are updated to be compatible with the tokenizer model. Updated unit tests (`test_awq.py` and `test_smoothquant.py`) address the above changes. Future plan: This PR doesn't address latency and consumption. We need to expand metrics for comprehensive benchmarks within AWQ/SmoothQuant.
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/3000
Note: Links to docs will display an error until the docs builds have been completed. This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
namgyu-youn
changed the title
namgyu-youn
changed the title
jerryzh168
approved these changes
Oct 8, 2025
namgyu-youn
changed the title
|
Removed the tokenizer because too many errors made AWQ api reproducible, mostly dispatch errors; just testing real models without |
namgyu-youn
changed the title
namgyu-youn
changed the title
namgyu-youn
changed the title
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Outdated suggestions cannot be applied.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.
Summary:
This is a milestone for building a benchmark for AWQ/SmoothQuant within the vLLM ecosystem (#2815). For vLLM benchmark, AWQ/SmoothQuant needs a wrapper (TransformerEvalWrapper) to be compatible with vLLM. Therefore, this PR adds unit test to verify if TransformerEvalWrapper works with AWQ/SmoothQuant.
Test plan:
A toy tokenizer model is introduced to validate TransformerEvalWrapper. Toy linear models are updated to be compatible with the tokenizer model.
Future plan:
This PR only validates the wrapper for the benchmark within the vLLM ecosystem. We need to expand metrics for comprehensive AWQ/SmoothQuant benchmarks.