Feynman27 / FPN

Feature Pyramid Network

FPN

---

end2end testing: mAP(0.7372) Further improvement may be done by decrease the BATCH_SIZE and learning rate for stability of loss curve

python ./faster_rcnn/test_net.py --gpu 0 --weights output/FPN_end2end/voc_0712_trainval/FPN_iter_300000.ckpt --imdb voc_0712_test --cfg ./experiments/cfgs/FPN_end2end.yml --network FPN_test

alt_opt training:

nohup ./experiments/scripts/FPN_alt_opt.sh 0 FPN_alt_opt pascal_voc0712 --set RNG_SEED 42 TRAIN.SCALES "[800]" > FPN_alt_opt.log 2>&1 &

end2end training:

nohup ./experiments/scripts/FPN_end2end.sh 1 FPN pascal_voc0712 --set RNG_SEED 42 TRAIN.SCALES "[800]" > FPN.log 2>&1 &

tail -f FPN.log

---

TODO:

1. imporve end2end training result, test alt_opt training result
2. acceleration by changing fake fpn_roi_pooling to real tf op