v0.7.2

Release xllm 0.7.2

Highlights

Feature

• Enhance Qwen3-MoE to support TP settings beyond 4.
• Implement chunked prefill and prefix cache for Qwen3 MoE.
• Support prefix cache for DeepSeek-V3/R1 models.

Bugfix

• Fix core dump issue triggered by client disconnection.
• Fix the incorrect reading of model args from Qwen3-VL's config.json.
• Setup the tokenizer config function of bos and eos to fast tokenizer.
• Fix the memory leak issue.
• Fix hang issue when enable schedule overlap.

Release Images

x86 image

quay.io/jd_xllm/xllm-ai:xllm-0.7.2-release-hb-rc2-x86

ARM a2 device image

quay.io/jd_xllm/xllm-ai:xllm-0.7.2-release-hb-rc2-arm

ARM a3 device image

quay.io/jd_xllm/xllm-ai:xllm-0.7.2-release-hc-rc2-arm