Support llama3 eagle3 head with llama4 verifier#117

Draft

rahul-tuli wants to merge 60 commits intomainfrom

support-llama3-eagle3-head-with-llama4-verifier

Commits on Oct 4, 2025

[cpu][perf] Accelerate unquantized-linear for AArch64 through oneDNN/ACL and weight prepack (#25948)

fadara01
and
bigPYJ1151
authored
[V1] [Hybrid] Mamba2 Automatic Prefix Caching (#25752)

authored
Support expert parallel in Transformers backend (#26162)

hmellor
and
Isotr0py
authored
[Model] Support nested structures for TensorSchema (#26212)
DarkLight1337
authored
[Misc] Require `merge_by_field_config` argument (#26214)
DarkLight1337
authored
[Misc] Remove unused `executor.apply_model` (#26215)
DarkLight1337
authored
[CI Failure] fix_test_auto_prefix_cache_support (#26053)
hl475
authored
Revert "Add batch invariant kernel override for FlashInfer backend [2/n]" (#26220)
DarkLight1337
authored
Add Olmo 3 reasoning parser (#26054)
soldni
authored
[Core] Enable decode of context length equal to max model length (#26168)
yannicks1
authored
[Bugfix] Fix `_reqs_to_process` leak on abort (#26012)
NickLucche
authored
[Model] CLIP Embedding Support (#26010)
DarkLight1337
authored
Fix tensor device and dtype placement in Qwen2VL model (#26219)

authored
[V1] [Hybrid] Remove code to override default CUDA graph configuration (#26226)
tdoublep
authored
[CPU] Refine batch reorder of CPU attention backend (#26096)
bigPYJ1151
authored
[Frontend] Cache chat template kwargs resolution (#26227)
Isotr0py
authored
[Renderer] Clean up renderer code (#26216)
DarkLight1337
authored

Commits on Oct 5, 2025

Commits on Oct 6, 2025