Closed marcovwu closed 3 years ago
Hi, Sorry for the delayed reply. As you have noted, you will need to decrease the learning rate when reducing the batch size. Do you also encounter the same error when doing so for the CenterNet2 model?
Thanks!! I tried to load pretrained model ( CenterNet2_R2-101-DCN-BiFPN_4x+4x_1560_ST.pth ), which sloved my problem.
Hi,i want to use CenterNet2_R2-101-DCN-BiFPN_4x+4x_1560_ST model to demo ,but i can't find the file of Base-CenterNet2.yaml.how to download the file.
I trained : python ./tools/train_net.py --config-file ./configs/COCO-Detection/faster_rcnn_R_50_C4_1x.yaml --num-gpus 1 SOLVER.IMS_PER_BATCH 2 SOLVER.BASE_LR 0.0025
Susccessful!!
But I run :
python ./projects/CenterNet2/train_net.py --config-file ./projects/CenterNet2/configs/CenterNet2_R2-101-DCN-BiFPN_4x+4x_1560_ST.yaml --num-gpus 1 SOLVER.IMS_PER_BATCH 2