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, 22:45 UTC
main
[Bugfix][ROCm] Fix load issue on deepseek quark quantization when shared expert enabled (#31261)
Pleaplusone committed 11 months ago
[CI/Build] Ignore data_parallel_size_local (#30281)
Ryan Rock committed 11 months ago
[ROCm][CI] Set TORCH_NCCL_BLOCKING_WAIT Distributed Tests On ROCm (#31259)
Micah Williamson committed 11 months ago
[ROCm][CI] Fix "Distributed Tests (H200)" Test (#31227)
kliuae committed 11 months ago
[ROCm][CI] Fix entrypoints tests and Python-only installation test on ROCm (#28979)
Andreas Karatzas committed 11 months ago
Add `--max-model-len auto` to auto-fit context to available memory (#29431)
Michael Goin committed 11 months ago
[docker] Fix downloading sccache on aarch64 platform (#30070)
Nick Cao committed 11 months ago
[XPU] Remove distributed_executor_backend check (#30760)
sihao_li committed 11 months ago
[Qwen3-Omni] fixed _get_feat_extract_output_lengths function (#31007)
Xiong Wang committed 11 months ago
[DeepSeek v3.2] Remove unnecessary syncwarps (#31047)
Matthew Bonanni committed 11 months ago
zejunchen-zejun committed 11 months ago
Revert "[bench] Support common prefix len config (for decode-only bench)" (#31240)
Ming Yang committed 11 months ago
[ROCm][CI] Set VLLM_FLOAT32_MATMUL_PRECISION="tf32" For terratorch Tests In AMD CI (#31242)
Micah Williamson committed 11 months ago
[ROCm][CI][Bugfix] Fix Siglip2 rotary embedding dispatch and InternVL video test tolerance (#31235)
Andreas Karatzas committed 11 months ago
[Bug] Fix `Number of dimensions of tensors must match.` for Deepseek V3.2 (#31160)
Wentao Ye committed 11 months ago
[P/D] Mooncake connector support more protocols (#30133)
Chao Lei committed 11 months ago
[KVEvent] User request.block_hash for parent block_hash (#30544)
Chen Zhang committed 11 months ago
[Misc] Remove unused custom ops `copy_blocks` and `copy_blocks_mla` (#30967)
rongfu.leng committed 11 months ago
[Bugfix] Enable `dynamic_dims` for different embeds shape (#31223)
Cyrus Leung committed 11 months ago
[Chore] Simplify logic of `_execute_mm_encoder` (#31222)
Cyrus Leung committed 11 months ago