FEAT: [Model] Support DeepSeek-V3.1 Quantization and tool

[Jun-Howie]

# ❗there are glitches with vllm 0.10.1.1, still looking for resolutions❗
# ❗downgrade vllm for now ❗
pip install vllm==0.9.2 transformers==4.53.0

SITE_PACKAGES=$(pip -V | awk '{print $4}' | sed 's/\/pip$//')
# ❗patch up AWQ MoE quant config, otherwise some modules cannot be properly loaded❗
cp awq_marlin.py "$SITE_PACKAGES/vllm/model_executor/layers/quantization/awq_marlin.py"
# ❗patch up for fp32 e_score_correction_bias, see https://www.github.com/vllm-project/vllm/pull/23640❗
cp deepseek_v2.py "$SITE_PACKAGES/vllm/model_executor/models/deepseek_v2.py"

[qinxuye]

LGTM