GNN 5 fold実行手順

[matsuken92]

対象インスタンス

• pytorch-gpu-002
• pytorch-gpu-003
• pytorch-gpu-004
• pytorch-gpu-005
• pytorch-gpu-006

gcloud compute ssh pytorch-gpu-002 --zone us-west1-b gcloud compute ssh pytorch-gpu-003 --zone us-west1-b gcloud compute ssh pytorch-gpu-004 --zone us-west1-b gcloud compute ssh pytorch-gpu-005 --zone us-west1-b gcloud compute ssh pytorch-gpu-006 --zone us-west1-b

データ

scp -i ~/key/google_compute_engine -r 10.138.0.23:/home/kenichi.matsui/work/molecular/input/graph4.zip ./

work/molecular/input/split/split_fold_0 work/molecular/input/split/split_fold_1 work/molecular/input/split/split_fold_2 work/molecular/input/split/split_fold_3 work/molecular/input/split/split_fold_4

work/molecular/input/node_std.npy work/molecular/input/node_mean.npy work/molecular/input/edge_std.npy work/molecular/input/edge_mean.npy

code

ソースコードをローカルからコピー

実行

# pytorch-gpu-002
nohup python -u train.py 0 &
tail -f nohup.out

# pytorch-gpu-003
nohup python -u train.py 1 &
tail -f nohup.out

# pytorch-gpu-004
nohup python -u train.py 2 &
tail -f nohup.out

# pytorch-gpu-005
nohup python -u train.py 3 &
tail -f nohup.out

# pytorch-gpu-006
nohup python -u train.py 4 &
tail -f nohup.out

gcloud compute ssh pytorch-gpu-007 --zone us-west1-b

pytorch-gpu-007

nohup python -u train.py 0 & tail -f nohup.out