gaoyang07 483bd706dd fix when resuming lr_scheduler without loading optimizer		2023-12-29 15:15:16 +08:00
..
apis	feat(tools): support origin internlm architecture in web_demo (#478)	2023-11-09 20:01:55 +08:00
core	fix(pp): fix no-packed dataset load micro batch error (#538)	2023-12-13 14:48:32 +08:00
data	fix the type_ids when micro_num=1 and use_flash_attn=False (#516)	2023-12-06 14:38:28 +08:00
initialize	feat(grad_norm): vocab grad norm profiling (#519)	2023-12-06 13:52:42 +08:00
model	fix(model): change model_type `LLAMA` to `LLAMA2` (#539)	2023-12-13 17:24:45 +08:00
moe	fix(moe): remove norm&gate force sync (#448)	2023-11-01 11:29:55 +08:00
monitor	fix(alert): send exception of all ranks (#491)	2023-11-10 19:04:31 +08:00
solver	fix when resuming lr_scheduler without loading optimizer	2023-12-29 15:15:16 +08:00
train	fix the type_ids when micro_num=1 and use_flash_attn=False (#516)	2023-12-06 14:38:28 +08:00
utils	fix when resuming lr_scheduler without loading optimizer	2023-12-29 15:15:16 +08:00
__init__.py	initial commit	2023-07-06 12:55:23 +08:00