Latest ResNet classifier for classification task

1. Unzip them and put the folders under ./checkpoints
2. Remember to set option -g to -g 0, -g -1 or -g mps

• Test ResNet18 on Tiny-ImageNet (Correctness: 7162/10000)

python test.py -d ./data/tiny-imagenet -n Resnet18onImageNet -g mps -m res_class --dataset_name imagenet --phase val --eval --net_name resnet18 --batch_size 4 --epoch best

• Test ResNet50 on Tiny-ImageNet (Correctness: 8196/10000)

python test.py --config config/ResNet50onTinyImageNet/test.yaml

• Test ResNet101 on Tiny-ImageNet (Correctness: 8602/10000)

python test.py -d ./data/tiny-imagenet -n Resnet101onImageNet -g mps -m res_class --dataset_name imagenet --phase val --eval --net_name resnet101 --batch_size 4 --epoch best

• Test ResNet50 on PASCAL_VOC_2012 (Correctness: 4022/5823)

python test.py -d ./data/pascal_voc_2012 -n Resnet50onPASCAL -g mps -m res_class --dataset_name pascalvoc2012 --phase val --eval --net_name resnet50 --batch_size 4 --epoch best  --loss_type bcewithlogits

• Test ResNet101 on PASCAL_VOC_2012 (Correctness: 4080/5823)

python test.py -d ./data/pascal_voc_2012 -n Resnet101onPASCAL -g mps -m res_class --dataset_name pascalvoc2012 --phase val --eval --net_name resnet101 --batch_size 4 --epoch best --loss_type bcewithlogits

• Test CNN on Brain Tumor MRI (Correctness: 1284/1311)

python test.py --config config/CNNonBrainMRI/test.yaml