yzwu 901b38c936 [Iluvatar] Optimize decode group_gemm and Support cuda graph for ernie (#6803)

2026-03-12 19:21:17 +08:00

..

graph_optimization

[Speculative Decoding] Unify Spec and non-spec branch (#6685)

2026-03-10 23:58:44 -07:00

guided_decoding

[Feature] Guided Decoding add LLguidance backend (#5124)

2025-12-03 20:23:57 +08:00

layers

[Feature] use phi permute/unpermute & rm swiglu (#6361)

2026-03-12 02:01:57 -07:00

logits_processor

[Feature] Support ThinkingBudget Logits processor to control thinking content length (#6367)

2026-02-25 14:17:09 +08:00

model_loader

add reconstruct (#6675)

2026-03-10 11:25:37 +08:00

models

[Feature] use phi permute/unpermute & rm swiglu (#6361)

2026-03-12 02:01:57 -07:00

ops

[Iluvatar] Optimize decode group_gemm and Support cuda graph for ernie (#6803)

2026-03-12 19:21:17 +08:00

__init__.py

…

entropy_utils.py

[Bugfix] Fix entropy calculation bugs (#5941)

2026-01-08 20:57:35 +08:00

forward_meta.py

[Models][OP][Optimization] Support DeepSeek-v3.2 model, integrate DSA & Indexer architecture with FlashMLA/DeepGEMM (#6689)

2026-03-10 15:05:14 +08:00

load_weight_utils.py

add reconstruct (#6675)

2026-03-10 11:25:37 +08:00

pre_and_post_process.py

[RL][Cherry-Pick] Support Fully Async and PrefixCache (#6599)

2026-03-12 01:13:30 -07:00

utils.py

[Others]update paddleformer 1.0.0 (#6496)

2026-03-11 15:06:29 +08:00

xpu_pre_and_post_process.py

[XPU] rm stop nums (#6651)

2026-03-12 14:05:58 +08:00