Netanel HaberandJee Jee Li 12a3f6454b [Bugfix][MoE] Only unpad routed output before shared expert add or routed output transform (#40865) ...

Signed-off-by: Netanel Haber <58652339+netanel-haber@users.noreply.github.com>
Co-authored-by: Jee Jee Li <pandaleefree@gmail.com>

2026-04-25 20:50:12 +00:00

..

kernels

Properly enable wvSplitK fp8 path for RDNA (#37712)

2026-04-20 10:09:24 -05:00

layers

[Bugfix][MoE] Only unpad routed output before shared expert add or routed output transform (#40865)

2026-04-25 20:50:12 +00:00

model_loader

[Model] Support Hy3 preview (#40681)

2026-04-23 22:08:26 +08:00

models

[Opt] Optimize deepstack buffer handling for multimodal Qwen3 models (#40145)

2026-04-25 21:04:26 +08:00

offloader

[Refactor] Clean up log once scope="local" (#40540)

2026-04-22 16:42:43 -04:00

warmup

mxfp8 online quant move to new frontend (#40152)

2026-04-20 06:26:12 -07:00

__init__.py

[Platform] Deprecate seed_everything (#31659)

2026-01-04 18:34:04 -08:00

custom_op.py

Add ability to replace oot ops when using lora (#37181)

2026-03-16 18:04:15 -07:00

parameter.py

[Quantization] add humming quantization kernel (#34556)

2026-04-24 09:29:44 -04:00

utils.py

[BugFix] Fix EPLB fail for MoeFP4 model with Marlin backend (#33262)

2026-01-29 16:52:11 +08:00