GoldPancake 4e10ae5d99 [Speculative Decoding] Optimize draft logprob (#5842) ...

* optimize draft logprob

* fix ut

2025-12-31 13:35:56 +08:00

..

graph_optimization

[Others] add assert and only count the actual load in cuda_graph (#5445)

2025-12-10 11:22:54 +08:00

guided_decoding

…

layers

[Speculative Decoding] Optimize draft logprob (#5842)

2025-12-31 13:35:56 +08:00

logits_processor

…

model_loader

[Loader]Fix bug in MTP weight loading (#5744)

2025-12-25 11:32:17 +08:00

models

[Feature] support w4afp8 v1_loader and v0_loader(tp>1) (#5757)

2025-12-30 14:11:52 +08:00

ops

[Iluvatar] Fix FD launch error when specifing CUDA_VISBLE_DEVICE (#5735)

2025-12-26 14:01:27 +08:00

__init__.py

…

entropy_utils.py

[BugFix] Fix entropy bugs (#5818)

2025-12-29 20:44:29 -08:00

forward_meta.py

[Intel HPU] enable tensor_wise_fp8 (#5324)

2025-12-17 16:45:03 +08:00

load_weight_utils.py

remove fastsafetensors (#5371)

2025-12-04 19:22:04 +08:00

pre_and_post_process.py

[Feature] Entropy calculation support (#5692)

2025-12-23 21:19:47 +08:00

utils.py

[Feature] support w4afp8 v1_loader and v0_loader(tp>1) (#5757)

2025-12-30 14:11:52 +08:00

xpu_pre_and_post_process.py

[XPU]ZMQ logprob (#5628)

2025-12-25 14:50:01 +08:00