A high-throughput and memory-efficient inference and serving engine for LLMs
main
[ROCm][CI]Fix test_concat_and_cache_mla_rope_fused on ROCm (#46409)
Divakar Verma committed 4 months ago
[MoE Backend] add HPC-Ops MoE backend (#45924)
Cheng Jiang committed 4 months ago
[ROCm][CI TG] refactor and fix deepep_moe test group (#46758)
Divakar Verma committed 4 months ago
[ROCm][CI] Relax fused layernorm quant test tolerances for one-ULP outliers (#46658)
Divakar Verma committed 4 months ago
fix(docker): eliminate race conditions in shared buildkit cache mounts (#44984)
weizhoublue committed 4 months ago
[Build] Show error message when using ROCm with LTO and different compilers (#35232)
Dāvis committed 4 months ago
Chris Leonard committed 4 months ago
[ModelRunner V2] Support realtime embeddings (#46762)
Nick Hill committed 4 months ago
Brandon Pelfrey committed 4 months ago
[CI] Don't try and download files that we already know don't exist (#46854)
Harry Mellor committed 4 months ago
[ModelRunner V2] Fix cross-attention block table sizing (#46753)
Nick Hill committed 4 months ago
[ModelRunner V2] Deduplicate ModelState init logic (#46776)
Nick Hill committed 4 months ago
Fix Transformers backend FP8 MoE and remove some boilerplate (#46820)
Harry Mellor committed 4 months ago
[ROCm][Bugfix] Pass num_kv_splits to aiter mla_reduce_v1 (#46760)
Rohan Potdar committed 4 months ago
[CI] Add @ivanium to CODEOWNERS for KV-cache/offload areas (#46873)
Yifan Qiao committed 4 months ago
[Frontend][Gpt-oss] Use `process_eos()` to flush Harmony Parser outputs. (#46437)
yzong-rh committed 4 months ago
[Core] Remove FlashAttention block size restriction for hybrid models (#36701)
Thomas Parnell committed 4 months ago
[GLM-5] Add DSV3.2/GLM5 to `vllm/models/` (#46808)
Woosuk Kwon committed 4 months ago
[ROCm][CI] Fix rlhf_nccl.py on ROCm (#46851)
Charlie Fu committed 4 months ago
[ROCm][CI] Remove V1 Sample + Logits from mi250 Queue (#46867)
Micah Williamson committed 4 months ago