For example, if you want to train the following models with 2 GPUs:
CUDA_VISIBLE_DEVICES=0,1 sh dist_train.sh 2 --cfg_file tools/cfgs/voxel/semantic_kitti/minkunet_mk34_cr10.yaml
CUDA_VISIBLE_DEVICES=2,3 sh dist_train.sh 2 --cfg_file tools/cfgs/voxel/semantic_kitti/cylinder_cy480_cr10.yaml
CUDA_VISIBLE_DEVICES=4,5 sh dist_train.sh 2 --cfg_file tools/cfgs/fusion/semantic_kitti/spvcnn_mk18_cr10.yaml
CUDA_VISIBLE_DEVICES=6,7 sh dist_train.sh 2 --cfg_file tools/cfgs/fusion/semantic_kitti/rpvnet_mk18_cr10.yaml
For example, if you want to train the following models with 2 GPUs:
CUDA_VISIBLE_DEVICES=0,1 sh dist_train.sh 2 --cfg_file tools/cfgs/voxel/waymo/minkunet_mk34_cr16.yaml
CUDA_VISIBLE_DEVICES=2,3 sh dist_train.sh 2 --cfg_file tools/cfgs/voxel/waymo/cylinder_cy480_cr10.yaml