Lan_se&shanzhu comments

Repositories
Issues
Comments

Results 2 comments of


                                            Lan_se&shanzhu

选择四卡训练卡住

> 使用二卡训练正常，大于2的都会只加载2次模型，然后卡住，不进行训练你好，我想请问下，使用2卡训练报错是怎么回事，我的指令式是这个NPROC_PER_NODE=2 xtuner train test_myllama_train.py --deepspeed deepspeed_zero2；显示torch.distributed.elastic.multiprocessing.errors.ChildFailedError:

xtuner如何根据checkpoints继续预训练

''' # whether to resume training from the loaded checkpoint resume = False ''' change 'False' into 'True'