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 14, 05:30 UTC
main
[Deprecation] Remove `prompt_token_ids` arg fallback in `LLM.generate` and `LLM.embed` (#18800)
Cyrus Leung committed about 1 year ago
[Kernel] Add FP8 support with FlashMLA backend (#22668)
Matthew Bonanni committed about 1 year ago
[Core] Support custom executor qualname (#23314)
22quinn committed about 1 year ago
[Feature][Responses API] Support logprobs(non-stream) (#23319)
Kebe committed about 1 year ago
tvalentyn committed about 1 year ago
[Perf] Small optimizations for silu_mul_fp8_quant_deep_gemm (#23265)
Michael Goin committed about 1 year ago
[CI] improve pr comments bot (#23380)
Simon Mo committed about 1 year ago
[CI] Clean up actions: remove helm, publish workflows and improve pr … (#23377)
Simon Mo committed about 1 year ago
[Structured Outputs] Refactor bitmask construction into get_grammar_bitmask (#23361)
Woosuk Kwon committed about 1 year ago
Make sure that vectorize_with_alignment produced vectorized global loads (#23182)
Elvir Crnčević committed about 1 year ago
[Core] Always use tensor cores for Flashinfer Decode Wrapper (#23214)
Pavani Majety committed about 1 year ago
[Misc] Convert VLLM_TORCH_PROFILER_DIR path to absolute (#23191)
Ning Xie committed about 1 year ago
[BugFix][gpt-oss] Fix Chat Completion with Multiple Output Message (#23318)
Chen Zhang committed about 1 year ago
[Misc] Misc code cleanup/simplification (#23304)
Nick Hill committed about 1 year ago
[Bugfix] Fix port conflict by obtaining a list of open ports upfront (#21894)
Ming Yang committed about 1 year ago
[Bug] Fix R1 Accuracy 0 Bug (#23294)
Wentao Ye committed about 1 year ago
add tg-mxfp4-moe-test (#22540)
Lain committed about 1 year ago
Yi Liu committed about 1 year ago
Migrate MolmoImageInputs to TensorSchema (#22022)
Benji Beck committed about 1 year ago
[ci/build] Fix abi tag for aarch64 (#23329)
youkaichao committed about 1 year ago