JimmyPeilinLi 168e10f254 [fix](sft): align Python API with C++ backend after v5 refactor ...

- wrapper.py: pass gpu_experts_mask=None to KTMoEWrapper (required by C++ signature)
- layer.py: rename submit_forward_sft/sync_forward_sft to submit_forward/sync_forward
- autograd.py: rename sync_forward_sft to sync_forward

The sft-v5 refactor (commits 58d7eab, dd1da65) renamed Python-side method
calls but the C++ backend (AMXSFTMoEWrapper) still exposes the original
method names. This caused AttributeError on Qwen3.5-35B and other models.

2026-04-20 16:44:09 +00:00

..

__init__.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

amx.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

arch.py

feat(sft): add Qwen3.5 MoE support + fused checkpoint loading

2026-04-20 17:19:15 +08:00

autograd.py

[fix](sft): align Python API with C++ backend after v5 refactor

2026-04-20 16:44:09 +00:00

base.py

feat(sft): AMX MoE SFT backend with LoRA support

2026-04-08 23:11:00 +08:00

config.py

refactor(sft): share_backward_bb default True, share_cache_pool auto-derived

2026-04-09 20:10:38 +08:00

dist_utils.py

refactor(sft): unify KTConfig field names with kt_ prefix, add share_cache_pool, remove dead code

2026-04-09 14:17:50 +08:00

layer.py

[fix](sft): align Python API with C++ backend after v5 refactor

2026-04-20 16:44:09 +00:00

lora.py

feat(sft): support transformers v5 fused expert format

2026-04-20 13:21:29 +08:00

weights.py

feat(sft): add Qwen3.5 MoE support + fused checkpoint loading

2026-04-20 17:19:15 +08:00

wrapper.py

[fix](sft): align Python API with C++ backend after v5 refactor

2026-04-20 16:44:09 +00:00