[Chore]: Stream tokens vs characters in tool call parser tests (#26513)

Ben Browning committed about 1 year ago

fixing mm placeholder replacement issue with gemma3 (#27538)

tingtinggithub committed about 1 year ago

[Bugfix][LoRA][FusedMoE] Select MxFP4 Backend based on LoRA Enablement (#27487)

Varun Sundar Rabindranath committed about 1 year ago

[Model] Siglip2 Model Support (#27566)

Yu Jiaqi committed about 1 year ago

[Hybrid] Added supports_mamba_prefix_caching Protocol (#27339)

Asaf Joseph Gardin committed about 1 year ago

[Kernel] Enable moe LoRA kernel support FP16 (#27468)

Jee Jee Li committed about 1 year ago

[cpu][perf] Fix low CPU utilization with VLLM_CPU_OMP_THREADS_BIND on AArch64 (#27415)

Fadi Arafeh committed about 1 year ago

[Bugfix] Fixed when return_token_ids=False, the first event still contains prompt_token_ids. (#27561)

Chauncey committed about 1 year ago

[Bugfix] Limit the default value of `max_model_len` when it is not specified by users (#27556)

Shanshan Shen committed about 1 year ago

[Misc] Clean up utils (#27552)

Cyrus Leung committed about 1 year ago

[Kernel] Adding split_K implementation for fused_moe_lora (#27291)

Danielle Robinson committed about 1 year ago

[Doc] Slight improvement to M2 and beyond (#27554)

Jee Jee Li committed about 1 year ago

[Model] Deprecate `merge_by_field_config=False` (#27551)

Cyrus Leung committed about 1 year ago

[Performance][LoRA] add context varying params to 'do_not_specialize' in fused moe lora (#27445)

gnovack committed about 1 year ago

[Docs] reemove the incorrect `enable_reasoning` parameter (#27550)

yyzxw committed about 1 year ago

[Model] Use merge_by_field_config for MM models (Qwen series) (#27546)

Cyrus Leung committed about 1 year ago

[Model][Bugfix] fix ernie45 moe 300B SharedFusedMoE output tuple (#27316)

CSWYF3634076 committed about 1 year ago

Fix MiniMax-M2 copyright (#27537)

Roger Young committed about 1 year ago

fix m2 test (#27536)

youkaichao committed about 1 year ago

[Model][MiniMax-M2] Support MiniMax-M2 Model (#27535)

Roger Young committed about 1 year ago