魅影 14abb75126 [model] enable using FA in npu (#9397) ...

Co-authored-by: frozenleaves <frozen@Mac.local>

2025-11-04 19:32:30 +08:00

..

__init__.py

rename files

2024-06-07 00:09:06 +08:00

attention.py

[model] enable using FA in npu (#9397)

2025-11-04 19:32:30 +08:00

checkpointing.py

[model] fix vit gradient checkpointing (#7830)

2025-04-23 22:48:48 +08:00

embedding.py

[misc] fix import error (#9299)

2025-10-17 17:46:27 +08:00

ktransformers.py

[train] KTransformers SFT as backend engine for LLaMA-Factory (#9400)

2025-11-04 15:54:12 +08:00

kv_cache.py

[misc] fix import error (#9299)

2025-10-17 17:46:27 +08:00

liger_kernel.py

[misc] update glm4v ligerkernel (#8978)

2025-08-20 23:39:56 +08:00

longlora.py

[trainer] fix pt loss (#7748)

2025-04-17 03:15:35 +08:00

misc.py

[misc] upgrade format to py39 (#7256)

2025-03-12 00:08:41 +08:00

mod.py

[misc] update license year & fix llama pro (#6814)

2025-02-05 01:53:33 +08:00

moe.py

[misc] fix import error (#9299)

2025-10-17 17:46:27 +08:00

packing.py

[trainer] fix pt loss (#7748)

2025-04-17 03:15:35 +08:00

quantization.py

[data] fix template (#8827)

2025-08-06 06:58:09 +08:00

rope.py

[model] switch to gptqmodel (#8108)

2025-05-19 22:25:40 +08:00

unsloth.py

Revert "[model] add lora dropout to unsloth" - requested feature already exists (#8554)

2025-07-05 11:25:31 +08:00

valuehead.py

Merge commit from fork

2025-06-26 13:55:42 +08:00

visual.py

[model] add deepstack_merger_list to Qwen3-VL vision_model_keys (#9399)

2025-11-04 19:27:34 +08:00