Yongye Zhu acb0f1dcdb [Bugfix][Spec Decode] DSpark: inherit the target's attention backend when the speculative config names none (#52288)

2026-08-15 09:32:59 +08:00

..

autoregressive

[Bugfix][MRV2] Require all requests to be decoding for uniform-decode dispatch (#51865)

2026-08-11 17:15:06 -07:00

dflash

Add Muse Glimmer model support (#51655)

2026-08-14 12:28:56 +08:00

dspark

[Bugfix][Spec Decode] DSpark: inherit the target's attention backend when the speculative config names none (#52288)

2026-08-15 09:32:59 +08:00

eagle

[Bugfix] DFlash fc sized wrong when num_target_layers != num_hidden_layers (#48524)

2026-07-21 14:42:52 -07:00

gemma4

[Kernel] Gemma-4 FA4 FP8 Kernel (#48666)

2026-08-13 17:42:00 -07:00

mtp

[Model Runner V2][MTP] Share topk index buffer between draft steps (#47352)

2026-08-10 15:46:13 -07:00

multi_module_mtp

[Bugfix][MRV2] Require all requests to be decoding for uniform-decode dispatch (#51865)

2026-08-11 17:15:06 -07:00

__init__.py

[Model Runner V2][Spec Decode] Add KV cache support for multi-layer MTP (#50062)

2026-08-13 16:22:29 -07:00

adaptive_verification.py

[Spec Decode] DSpark confidence-scheduled verification (#47808)

2026-08-12 12:27:48 -07:00

rejection_sampler_utils.py

[Model Runner V2] Fix -1 placeholder draft token ids in rejection sam… (#50939)

2026-08-06 16:32:00 -07:00

rejection_sampler.py

[Spec Decode] DSpark confidence-scheduled verification (#47808)

2026-08-12 12:27:48 -07:00

speculator.py

[Model Runner V2] Cache draft logits in model's LM head dtype (#50910)

2026-08-05 19:12:46 -07:00

utils.py

[Perf] Use blocking CUDA events to avoid busy polling cuda driver lock (#47081)

2026-07-07 09:36:10 +00:00