vLLM v0.11.0
Highlights This release features 538 commits, 207 contributors (65 new contributors)! This release completes the removal of V0 engine. V0 engine code including AsyncLLMEngine, LLMEngine, MQLLMEngine, all attention…
1 stories · 7d·6 sources covering·3 active storylines
What this is
DeepSeek-V3 is DeepSeek's open-weight mixture-of-experts base model. shipfeed tracks DeepSeek-V3 releases, weights, and the fine-tunes and benchmarks built on it.
Highlights This release features 538 commits, 207 contributors (65 new contributors)! This release completes the removal of V0 engine. V0 engine code including AsyncLLMEngine, LLMEngine, MQLLMEngine, all attention…
DeepSeek released V3.1, an optimized model designed for improved speed and better performance on Chinese-made hardware.
llama : MTP support for DeepSeek V3.2 (#26457) llama : MTP support for DeepSeek V3.2 model : no need to include MTP layers during DeepSeek V3.2 model type discovery --------- Co-authored-by: Stanisław Szymczyk Website…