MingkunZhang c4abb01f9c [Metax][Fix] fix 'get_token_penalty_multi_scores' input error based (PaddlePaddle#6069) (#6266)

2026-01-29 19:24:36 +08:00

..

attention

[Others] enhance deep_ep import and support mixed mode flash_mask_attn (#6238)

2026-01-28 00:02:02 +08:00

backends

[XPU] change XPU EP interface from xDeepEP to paddle (#5706)

2026-01-21 18:23:45 +08:00

batch_invariant_ops

…

moe

[Feature] Support NVFP4 MoE on SM100 (#6003)

2026-01-29 14:16:07 +08:00

pool

…

quantization

[Feature] Support NVFP4 MoE on SM100 (#6003)

2026-01-29 14:16:07 +08:00

sample

[Metax][Fix] fix 'get_token_penalty_multi_scores' input error based (PaddlePaddle#6069) (#6266)

2026-01-29 19:24:36 +08:00

__init__.py

…

activation.py

…

embeddings.py

…

linear.py

[Models][BugFix] shared experts and dense mlp layer do not require TP split (#6180)

2026-01-28 18:58:19 +08:00

lm_head.py

…

mtp_linear.py

…

normalization.py

Support MXFP4 for GPT-OSS (#5435)

2026-01-22 14:21:01 +08:00

pooler.py

…

rotary_embedding.py

…

utils.py

[Feature] Support Ernie FP8 on sm100 (#5593)

2026-01-29 13:49:54 +08:00