A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 10, 05:15 UTC
main
[Refactor] [6/N] to simplify the vLLM openai chat_completion serving architecture (#32240)
Chauncey committed 10 months ago
[Quantization] fix: overflow with static per-tensor scaling (#29867)
Mickaël Seznec committed 10 months ago
[Docs] Nixl Usage recommend `fail` kv_load_failure_policy (#32198)
Nicolò Lucchesi committed 10 months ago
[Bugfix] Replace `PoolingParams.normalize` with `use_activation` (#32243)
Cyrus Leung committed 10 months ago
[Refactor] Remove `get_encoder_dummy_data` (#32241)
Cyrus Leung committed 10 months ago
[Model] Use mm_position to compute mrope positions for Qwen2-VL/2.5-VL (#32126)
YunzhuLu committed 10 months ago
[ROCm][CI] Fix engine core client tests for ROCm spawn multiprocessing (#32061)
Andreas Karatzas committed 10 months ago
[Doc] Update installation from source command (#32239)
Roy Wang committed 10 months ago
[BugFix]Fix eagle draft_model_config and add tests (#31753)
Xingyu Liu committed 10 months ago
[ROCm][CI] Fix HuggingFace flash_attention_2 accuracy issue in Isaac vision encoder (#32233)
Andreas Karatzas committed 10 months ago
[ROCm][Bugfix] Fix Mamba batched decode producing incorrect output (#32099)
Andreas Karatzas committed 10 months ago
[Perf] Optimize requests abort (#32211)
Wentao Ye committed 10 months ago
Fix various typos found in `docs` (#32212)
Andrew Bennett committed 10 months ago
[Frontend] Add `reasoning_effort` to `OpenAIServing._preprocess_chat()` (#31956)
Sanghoon Yoon committed 10 months ago
[Misc] improve warning/assert messages (#32226)
cjackal committed 10 months ago
[BugFix] Fix engine crash caused by chat tools + response_format (#32127)
Nick Hill committed 10 months ago
[Misc] Allow enabling NCCL for DP sync when async scheduling (#32197)
Nick Hill committed 10 months ago
[Model] Handle `trust_remote_code` for transformers backend (#32194)
Cyrus Leung committed 10 months ago
[responsesAPI] add unit test for optional function tool call id (#32036)
Andrew Xia committed 10 months ago
[ROCm][CI] Handle pytest status code 5 when a shard isn't allocated any tests (#32040)
Divakar Verma committed 10 months ago