search

lxtGH / OMG-Seg

OMG-LLaVA and OMG-Seg codebase [CVPR-24 and NeurIPS-24]
Other
1.32k stars 50 forks source link

Multi node training? #35

Closed echo840 closed 2 months ago

echo840 commented 3 months ago

Does this code support multi-machine training, and how should I modify the training command?

HarborYuan commented 2 months ago

You do not need to modify the code. If you are using Slurm, you should use this script:

https://github.com/lxtGH/OMG-Seg/blob/main/tools/slurm.sh

Else you are expected to use this:

https://github.com/lxtGH/OMG-Seg/blob/main/tools/dist.sh

In this case, you need to set the MASTER_ADDR, PORT, NODE_RANK, NNODES, and GPUS by yourself.