bash.sh

################## pre-train stage ##################
## get the pre-train model in ./output
mv 'merge_domain_8/merge_0.json' '/cache/data/haitian/annotations/instance_unlabel_0.json'      ## pretrain on labeled domain
python train_net.py --num-gpus 8 --config configs/haitian_supervision/faster_rcnn_R_50_FPN_sup_run1.yaml SOLVER.IMG_PER_BATCH_LABEL 16 SOLVER.IMG_PER_BATCH_UNLABEL 16 SEMISUPNET.PARA_MU 0.1 SEMISUPNET.PARA_T 0.7

################## Evaluation on Unlabeled data ##################
## get the Evaluation results in ./output/infrence/coco_instances_results.json
python train_net.py --num-gpus 8 --eval-only --config configs/haitian_supervision/faster_rcnn_R_50_FPN_sup_run1.yaml SOLVER.IMG_PER_BATCH_LABEL 16 SOLVER.IMG_PER_BATCH_UNLABEL 16

################## Get 1. Domain Similarity 2. Estimated Class Distribution ##################
## input : ./output/infrence/coco_instances_results.json
## output : Domain Similarity & Estimated Class Distribution
python get_domain_similarity_class_distribution.py

################## DucTeacher stage ##################
## after getting the domain similarity and class distribution, we can train the DucTeacher.
python DucTeacher_domain_evolve_sh_0_10.py