Skip to content

Can't use distributed processing #13

Description

@SaikaOfficial

Thank you for sharing the code!
This is my script run.sh:

CUDA_VISIBLE_DEVICES=0,1,2,3 OMP_NUM_THREADS=1 python -m torch.distributed.launch --nproc_per_node=4 --master_port 12345 main_pretrain.py \
    --num_workers 10 \
    --accum_iter 2 \
    --batch_size 128 \
    --model mrm \
    --norm_pix_loss \
    --mask_ratio 0.75 \
    --epochs 200 \
    --warmup_epochs 40 \
    --blr 1.5e-4 --weight_decay 0.05 \
    --resume ./MRM/mae_pretrain_vit_base.pth \
    --data_path ./MRM \
    --output_dir ./MRM \

When I use distributed computing, the program always gets stuck in this position

image

and never continue. But if I set the graphics card to one, it can train at a very slow speed.

I'm wondering how to deal with it.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions