vllm-project/vllm

Pull Requests Commits

Merge branch 'main' into m3_release

ywang96 committed 10 hours ago

Verified 0c6d468d

[ROCm][DSV4][Perf] Fuse inverse-RoPE and cache bf16 wo_a in o-projection (#45103)

Fangzhou-Ai committed 11 hours ago

Verified badddd25

Add the QuantizedActivation linear-kernel contract (#44260)

mgoin committed 11 hours ago

Verified c9065008

[Kernel] Consolidate Marlin thread-tile padding across all dense Marlin paths (#45295)

mgoin committed 11 hours ago

Verified 9eaacb23

[Model Runner v2] Migration from v1 to v2, with Qwen and DSv2 MOE models [3/N] (#42667)

yewentao256 committed 11 hours ago

Verified 78739c19

[Attention] Improve attention benchmarks: configs and profiling (#39336)

MatthewBonanni committed 11 hours ago

Verified cf567cbc

[ROCm] Bump Torch to 2.11 (#45362)

micah-wil committed 11 hours ago

Verified 39cb9bf2

[Refactor] Deprecate ResponsesParser wrapper, inline parsing into ParsableContext (#45431)

sfeng33 committed 11 hours ago

Verified 6e4a5471

[Core][AMD] Propagate shutdown timeout to MultiprocExecutor (#43154)

rjrock committed 11 hours ago

Verified aab639c7

[Perf] Use native DSA indexer decode path for next_n > 2 on SM100 (#45322)

zixi-qi committed 12 hours ago

Verified efe7adb5

[Migration] Migrate GGUF quantization support to plugin (#39612)

Isotr0py committed 12 hours ago

Verified 6635279d

[Model][Dflash] Enable Dflash support for Qwen3NextForCausalLM targets (#45319)

j-i-l committed 14 hours ago

Verified d6fd7ce8

[Kernel][Helion][1/N] Add Helion kernel for dynamic_per_token_scaled_fp8_quant (#33790)

xiaohongchen1991 committed 15 hours ago

Verified 272c1695

jeejeelee committed 15 hours ago

4a560dd8

Address conflict

jeejeelee committed 15 hours ago

634f41ea

[Model] Add encoder CUDA graph support to Lfm2VL (#44930)

vincentzed committed 15 hours ago

Verified 053e7daa

[Rust Frontend] Add standalone `granite4` tool parser (#45216)

tahsintunan committed 15 hours ago

Verified 5af4aec1

Merge branch 'main' into m3_release

jeejeelee committed 16 hours ago

Verified 59e521f3

[Docs][KV Connector][NIXL] document KV Transfer stat logging and Prometheus metrics (#44055)

sridhar-3009 committed 16 hours ago

Verified a30addc7

[Frontend] Support strict mode for tool calling with ResponsesAPI (#45396)

chaunceyjiang committed 17 hours ago

Verified 3b8fc3fe

[Core][KV Connector] fix scheduler KV connector stats aggregation (#43877)

Srinivasoo7 committed 17 hours ago

Verified 9ff278b1

[Core] Support structured outputs for beam search (#35022)

guan404ming committed 18 hours ago

Verified c7aa3d26

relax constraint

gau-nernst committed 18 hours ago

Verified bd546f10

[Bug] Migrate Reset cache for both v2 and v1 model runner (#42759)

yewentao256 committed 18 hours ago

Verified fbc3a190

[V1][Metrics] Add MLA attention metrics for DeepSeek MFU estimation (#39457)

thillai-c committed 18 hours ago

Verified 4171ae40

minor fix for tool parser tests

BugenZhao committed 18 hours ago

Verified c1214b98

fix kernels tests import error

Isotr0py committed 19 hours ago

ce695f81

fix failed basic models test

Isotr0py committed 20 hours ago

d6105026

[Metrics] Add group-aware KV cache capacity to vllm:cache_config_info (#42206)

chfeng-cs committed 20 hours ago

Verified b7f9b6ab

[BUGFIX][XPU] Update fa interface for compatibility (#45394)

zhenwei-intel committed 20 hours ago

Verified 8af550b3

Older