[CPU] Support CPU W4A16 INT4 MoE (#43409)

Yuwen Zhou committed 5 months ago

[Bugfix] Fix Dockerfile dependency graph pre-commit error (#45374)

Isotr0py committed 5 months ago

[ROCm][gpt-oss] Pass GateMode.INTERLEAVE for MXFP4 W4A16 fused MoE (#44893)

Rohan Potdar committed 5 months ago

[MM][Perf][CG] Support ViT full cudagraphs for mllama4 (#40660)

allgather committed 5 months ago

[Model] Add DiffusionGemma Support (#45163)

Lucas Wilkinson committed 5 months ago

[ROCm][CI] fix fp8 support for test_deepep_moe (#45302)

Divakar Verma committed 5 months ago

[NIXL] Per-region KV transfer classification for mixed full-attn + MLA groups (#44583)

Dao007forever committed 5 months ago

[11a/n] Migrate Marlin kernels to torch stable ABI (#45176)

Chris Leonard committed 5 months ago

[ASR] Add Long Audio benchmark and correctness test (#44587)

Ekagra Ranjan committed 5 months ago

[Bugfix] Mamba CPU Offloading (#44599)

Varun Sundar Rabindranath committed 5 months ago

[ASR] Optimize CPU preproc to get 2.5x RTFx via multi-threading (#44612)

Ekagra Ranjan committed 5 months ago

[Bugfix] Fix --enable-prompt-tokens-details omitting zero cached tokens (#44383)

sasindharan committed 5 months ago

[CI][BugFix] Fix broken `test_mamba_prefix_cache.py` due to stale mock (#45345)

Nick Hill committed 5 months ago

[Bugfix][Model] Pass revision by name in Run:ai and bitsandbytes index downloads (#45308)

Ting SUN committed 5 months ago

[ROCm][DSv4][Perf] Flash-decode split-K decode attention kernel (#44899)

Fangzhou Ai committed 5 months ago

[Bugfix] OffloadingConnector: respect skip_reading_prefix_cache flag (#44592)

littlecircle0730 committed 5 months ago

[Refactor] Chat Completions Streaming Harmony Refactor and Bugfixes (#45104)

yzong-rh committed 5 months ago

[Bugfix] Initialize missing attributes in mistral eagle (#45217)

jpwang committed 5 months ago

Make mistral_common optional by deferring MistralToolCall import (#45305)

Neil Schemenauer committed 5 months ago

[Bugfix][KVConnector][Mooncake] Close MooncakeDistributedStore on connector teardown (#45206)

Dao007forever committed 5 months ago