[Model][Mamba] Add selector for mamba attention backend and make it pluggable for other device (#26487)

Shanshan Shen committed 12 months ago

[BugFix] Fix async-scheduling + FlashAttn MLA (#28990)

Lucas Wilkinson committed 12 months ago

[torch.compile] caching of config fields should be opt-out by default (#26468)

vnadathur committed 12 months ago

[CI] Reorganize compile tests so new tests are automatically included in CI (#28625)

Yanan Cao committed 12 months ago

Relax Transformers modeling backend MoE experts check (#28952)

Harry Mellor committed 12 months ago

[Doc]: fix typos in various files (#29010)

Didier Durand committed 12 months ago

[Feat][Perf] Enable deepep-low-latency with round-robin expert placement. (#28449)

Chen Bruce committed 12 months ago

[config] Expose `get_total_num_hidden_layers()` in ModelConfig (#28961)

Tova Movshovitz committed 12 months ago

[Docs] Take env var definition out of folded admonition (#29005)

Harry Mellor committed 12 months ago

[CI/Build] Fix broken build on Apple M1 (#28999)

j20120307 committed 12 months ago

refactor(cpu_types_scalar.hpp): Unify scalar loop implementations using unroll_loop (#28847)

ihb2032 committed 12 months ago

[Docs] Clean up moe_kernel_features.md (#28530)

Michael Yao committed 12 months ago

add support for --fully-sharded-loras in fused_moe (#28761)

gnovack committed 12 months ago

Add CPU support model (#28697)

Louie Tsai committed 12 months ago

[Doc]: fix typos in various files (#28945)

Didier Durand committed 12 months ago

Speed up macOS smoke test (#28954)

Michael Goin committed 12 months ago

[Misc] Update embedding/cross encoder tests to use `mteb` v2 (#27329)

Roman Solomatin committed 12 months ago

[Model][QwenVL] Simplify cos/sin rotary embedding indexing (#28962)

Lukas Geiger committed 12 months ago

[Docs] Update oneshot imports (#28188)

Uranus committed 12 months ago

[Bugfix] Fix typo in Qwen3 Next model executor (#28960)

Gleb Kurchanov committed 12 months ago