[Example] Add Mixed W2A16 and W4A16 MoE Example

[yiliu30]

Signed-off-by: yiliu30 yi4.liu@intel.com

SUMMARY:
"please provide a brief summary"

• model https://huggingface.co/INC4AI/Qwen3-30B-A3B-Instruct-2507-Attn4bits-Mlp2bits-AutoRound-200iters-TestOnly

# vllm ({'pretrained': '/home/yiliu7/workspace/llmc-ds/examples/autoround/quantization_wNa16/Qwen3-30B-A3B-Instruct-2507-Attn4bits-Mlp2bits-AutoRound-200iters', 'tensor_parallel_size': 1, 'max_model_len': 8192, 'max_num_batched_tokens': 32768, 'max_num_seqs': 128, 'add_bos_token': True, 'gpu_memory_utilization': 0.4, 'dtype': 'bfloat16', 'max_gen_toks': 2048, 'enable_prefix_caching': False}), gen_kwargs: ({}), limit: 1000.0, num_fewshot: None, batch_size: 128
# |Tasks|Version|     Filter     |n-shot|  Metric   |   |Value|   |Stderr|
# |-----|------:|----------------|-----:|-----------|---|----:|---|-----:|
# |gsm8k|      3|flexible-extract|     5|exact_match|↑  |0.909|±  |0.0091|
# |     |       |strict-match    |     5|exact_match|↑  |0.909|±  |0.0091|

The evluation is depends on vllm's PR vllm-project/vllm#48918

[github-actions]

👋 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.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: a26e80aa-808b-4014-8cc4-834ddb17615e

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gemini-code-assist]

Code Review

This pull request introduces a new example script, qwen3_moe_example_mixed_w2a16_w4a16.py, demonstrating mixed-precision quantization (W2A16 and W4A16) using AutoRound on a Qwen MoE model. The feedback suggests three key improvements: replacing the non-public model ID with a publicly available one, loading the model with torch_dtype="auto" to prevent potential out-of-memory errors, and increasing the number of calibration iterations from 2 to a standard value like 200 to ensure high-quality quantization.

[mergify]

Merge Protections

🟢 All 2 merge protections satisfied — ready to merge.

Show 2 satisfied protections

🟢 Require one maintainer review

All PRs must have at least one approving review from a maintainer before merging.

• #changes-requested-reviews-by = 0
• any of:
  • approved-reviews-by=dsikka
  • approved-reviews-by=kylesayrs
  • approved-reviews-by=HDCharles
  • approved-reviews-by=brian-dellabetta
  • approved-reviews-by=yiliu30

🟢 Require two reviews

PRs labelled "two-reviews" must have at least two approving reviews before merging.

• #approved-reviews-by >= 2
• #changes-requested-reviews-by = 0

[mergify]

The quality checks have failed. Please run make style and make quality under
the root directory to adddress the lint failures. You will need to install the
dev optional install to get the required linting packages:
https://github.com/vllm-project/llm-compressor/blob/main/CONTRIBUTING.md