RuntimeError: CUDA out of memory

[coba0907]

During RCNN training, the following error occurred when I run the following command.
command: python train_rcnn.py --cfg_file cfgs/default.yaml --batch_size 2 --train_mode rcnn --epochs 70 --ckpt_save_interval 2 --rpn_ckpt ../output/rpn/default/ckpt/checkpoint_epoch_200.pth

RuntimeError: CUDA out of memory. Tried to allocate 512.— MiB (GPU 0; 2.95 GiB total capacity; 1.81 GiB already allocated; 356.00 MiB free; 1.89 GiB reserved in total by PyTorch )

The same Error occurred in RPN traing, which was solved by reducing batch_size to 2. but I don't think that reducing the batch_size anymore.

can you help me?