initial commit on compressed-tensors quantization support for fp8

[qihqi]

Description

This PR is to enable models with FusedMoE module (such as Qwen3-coder) to run with compressed-tensors quantization scheme.

NOTE:
currently because of old jax version, we need to change
this line /home/hanq_google_com/uv_venv/lib/python3.12/site-packages/jax/experimental/pallas/ops/tpu/megablox/common.py: 44
which is:

_TPU_KIND_PATTERN = re.compile(r"TPU v(\d+)")

To:

_TPU_KIND_PATTERN = re.compile(r"TPU.*(\d+)")

Tests

MODEL_IMPL_TYPE=vllm VLLM_DISABLE_SHARED_EXPERTS_STREAM=1 python examples/offline_inference.py --model=BCCard/Qwen3-Coder-480B-A35B-Instruct-FP8-Dynamic --tensor_parallel_size=8 --task=generate --max_model_len=128 --max_num_seqs=1

runs and produces good result.

[github-actions]

Description

Start with a short description of what the PR does and how this is a change from
the past.

The rest of the description includes relevant details and context, examples:

• why is this change being made,
• the problem being solved and any relevant context,
• why this is a good solution,
• some information about the specific implementation,
• shortcomings of the solution and possible future improvements.

If the change fixes a bug or a Github issue, please include a link, e.g.,:
FIXES: b/123456
FIXES: #123456

Tests

Please describe how you tested this change, and include any instructions and/or
commands to reproduce.

Checklist

Before submitting this PR, please make sure:

• I have performed a self-review of my code.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have made or will make corresponding changes to any relevant documentation.

[kyuyeunk]

I've noticed that the code doesn't have any unit test? Can you add one?

[qihqi]

I've noticed that the code doesn't have any unit test? Can you add one?

added unit test, PTAL. Thanks!