Squelch MLA warning for Compressed-Tensors Models

[kylesayrs]

Purpose

• Squelch MLA warning when loading compressed-tensors models

WARNING 02-03 14:14:32 config.py:1007] compressed-tensors MLA support requires fp8 activations and weights in group 'group
_0', but got activations type 'None' and weights type 'int'.
WARNING 02-03 14:14:32 config.py:1007]  Full config: {'config_groups': {'group_0': {'input_activations': None, 'output_act
ivations': None, 'targets': ['Linear'], 'weights': {'actorder': None, 'block_structure': None, 'dynamic': False, 'group_si
ze': 128, 'num_bits': 4, 'observer': 'mse', 'observer_kwargs': {}, 'strategy': 'group', 'symmetric': True, 'type': 'int'}}
}, 'format': 'marlin-24', 'global_compression_ratio': 1.8917232374233346, 'ignore': ['lm_head'], 'kv_cache_scheme': None, 
'quant_method': 'compressed-tensors', 'quantization_status': 'compressed', 'sparsity_config': {'format': 'dense', 'global_
sparsity': 0.4561021413812781, 'ignore': ['lm_head'], 'registry_requires_subclass': False, 'sparsity_structure': '2:4', 't
argets': ['Linear']}}

Changes

• Use demorgans to invert final condition and place at the beginning to act as a guard before warnings

Testing

from vllm import llm
llm = LLM("neuralmagic/Sparse-Llama-3.1-8B-ultrachat_200k-2of4-quantized.w4a16")

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can do one of these:

• Add ready label to the PR
• Enable auto-merge.

🚀

[LucasWilkinson]

LGTM! thanks for the quick catch and fix

[tlrmchlsmth]

Looks good and thanks for the fix

FYI @kylesayrs the PR description looked a little messed up (It was just a header that said Purpose), so I edited it to add a small description. Could you take a look to make sure it looks good?

[kylesayrs]

@tlrmchlsmth Description updated