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 8, 19:00 UTC
main
[Minor] Sort safetensors files to ensure deterministic loading order (#33491)
Lumosis committed 9 months ago
[Bugfix] Make MM batching more robust (#33817)
Cyrus Leung committed 9 months ago
[Bugfix] Fix DSV3.2 NVFP4 (#33932)
Matthew Bonanni committed 9 months ago
[Misc] Rename `translations` to `speech_to_text` for OAI serving component (#33904)
Nicolò Lucchesi committed 9 months ago
Fix tokenizer test for renamed attr on Transformers v5 (#33902)
Harry Mellor committed 9 months ago
[Bugfix] Suppress non-TTY color output on the process name part of the log (#29714)
Tsukasa OI committed 9 months ago
[Models] Consolidate Deepseek-OCR2 processor (#33909)
Isotr0py committed 9 months ago
[Moe Refactor] Make Inplace Flag for FusedMoEModularKernel part of the constructor (#33375)
bnellnm committed 9 months ago
[Bugfix] Fix swapped engine_ids in NIXL Llama 4 local attention path (#33795)
zackyoray committed 9 months ago
[Misc] Add debug logs (#33931)
Nicolò Lucchesi committed 9 months ago
[Spec Decode] Unified Parallel Drafting (#32887)
Benjamin Chislett committed 9 months ago
[BugFix] Fix LoRA Fp8 (#33879)
danisereb committed 9 months ago
[Feat][RL][1/2] Native Weight Syncing API: NCCL (#31943)
Aaron Hao committed 9 months ago
[Bugfix] Fix step3p5 parser when using mtp (#33690)
Mario Hong committed 9 months ago
[Docs] Add bart-plugin to docs (#33905)
Nicolò Lucchesi committed 9 months ago
[Bugfix] Fix corner case of sparse embedding (#33886)
wang.yuqi committed 9 months ago
[Refactor] Clean up input preprocessing (#33687)
Cyrus Leung committed 9 months ago
[Bugfix] Fix Kimi-K2.5 NVFP4 checkpoints weight loading (#33876)
Isotr0py committed 9 months ago
[perf] Integrate flashinfer concat_mla_k (#31171)
jiahanc committed 9 months ago
Enable Cross layers KV cache layout at NIXL Connector V2 (#33339)
liranschour committed 9 months ago