LiYa comments

Repositories
Issues
Comments

Results 3 comments of


                                            LiYa

Mistake when apply parallel test

The command > CUDA_VISIBLE_DEVICES=1,2 python tools/test_localizer.py ${CONFIG_FILE} ${CHECKPOINT_FILE} --out ${RESULT_FILE} --eval activitynet --gpus 2 The error log > Traceback (most recent call last): File "tools/test_localizer.py", line 215, in main() File...

TencentPretrain pretrain.py of llama-7b model, Exception: : Current loss scale already at minimum

same problem

AssertionError: Check batch related parameters. train_batch_size is not equal to micro_batch_per_gpu * gradient_acc_step * world_size

As for my case, I found that I forgot to set the parameter in the deepspeed_config to "auto"