kiroxu cdb8545a91 [Kernel][Perf] Support Qwen head ratios in fused GDN MTP (#52539) ...

Signed-off-by: kiroxu <148877251+BabyDrangoner@users.noreply.github.com>

2026-08-17 18:17:40 -07:00

..

hw_agnostic

Hardware-agnostic model definition via HF transformer backend (1/N) (#49458)

2026-08-13 11:21:50 +01:00

kernels

[Refactor] Simplify B12X linear kernels and warmup (#52368)

2026-08-17 09:45:45 -07:00

layers

[Kernel][Perf] Support Qwen head ratios in fused GDN MTP (#52539)

2026-08-17 18:17:40 -07:00

model_loader

[Bugfix] Avoid repeated layerwise reload warning scans (#49505)

2026-08-12 13:53:50 +08:00

models

[ModelRunnerV2] Support prompt embeds (#42963)

2026-08-17 21:58:50 +00:00

offloader

Revert "[Platform] Replace torch.cuda.Event with torch.Event (#47140)" (#47668)

2026-07-06 14:18:33 +01:00

warmup

[3/N][Feat][Perf] Add new warmup infrastructure for JITs. Add provider registry and orchestration for JIT warmup (#50174)

2026-08-17 10:11:14 -07:00

__init__.py

[Platform] Deprecate seed_everything (#31659)

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

custom_op.py

[Bugfix] Reject contradictory custom-op directives (#49134)

2026-07-26 08:42:14 -04:00

parameter.py

Speed up docs build (#44635)

2026-06-05 14:51:44 +00:00

utils.py

[Weight processing] Copy over new_data attributes in replace_parameter (#49601)

2026-08-06 13:03:37 -04:00