This is the test code of ensemble mask-aided r-cnn for EAD2019 challenge workshop
Pytorch==1.0.0
maskrcnn-benchmark==0.1
| ID | Model | Backbone | Weights |
|---|---|---|---|
| 0 | Faster R-CNN | Resnet50 | model_final.pth |
| 1 | Mask-aided R-CNN | Resnet50 | model_final.pth |
| 2 | Faster R-CNN | Resnet50+FPN | model_final.pth |
| 3 | Mask-aided R-CNN | Resnet50+FPN | model_final.pth |
| 4 | Faster R-CNN | Resnet101+FPN | model_final.pth |
| 5 | Mask-aided R-CNN | Resnet101+FPN | model_final.pth |
| 6 | Faster R-CNN | ResneXt101+FPN | model_final.pth |
| 7 | Mask-aided R-CNN | ResneXt101+FPN | model_final.pth |
python3.6 tools/test_net.py --config-file "configs/e2e_mask_rcnn_R_50_C4_1x_ead_whole.yaml" MODEL.MASK_ON True MODEL.ROI_BOX_HEAD.NUM_CLASSES 6 TEST.IMS_PER_BATCH 1 OUTPUT_DIR "./Checkpoint/01_resnet50_mask_decay_1e_4"
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent