Specifying gpu-ids with mim train about mim HOT 6 CLOSED

Hi @okotaku , the command mim train mmcls config.py --gpu-ids 1 --gpus 1 should work well because gpu-ids will be parsed as other_args and they will also be passed to train.py.

from mim.

Hi, okotaku. In mmcls, --gpu-ids and --gpus are two exclusive args. Maybe you can use CUDA_VISIBLE_DEVICES=1 mim train mmcls config.py --gpus 1 instead.

from mim.

Thank you!

from mim.

When I try to use mim train mmcls config.py --gpu-ids 1 --gpus 1 , I get the following error.

usage: train.py [-h] [--work-dir WORK_DIR] [--load-from LOAD_FROM] [--resume-from RESUME_FROM] [--no-validate] [--gpus GPUS | --gpu-ids GPU_IDS [GPU_IDS ...]] [--seed SEED]
                [--deterministic] [--options OPTIONS [OPTIONS ...]] [--cfg-options CFG_OPTIONS [CFG_OPTIONS ...]] [--launcher {none,pytorch,slurm,mpi}] [--local_rank LOCAL_RANK]
                config
train.py: error: argument --gpu-ids: not allowed with argument --gpus

from mim.

Please @kennymckormick have a look.

from mim.

Thank you!

from mim.