Faster R-CNN R-101-FPN model was implemented with TensorFlow2.0 Eager Execution.

• Cuda 10.0
• Python 3.5
• TensorFlow 2.0.0
• cv2

see train_model.ipynb, inspect_model.ipynb and eval_model.ipynb

• 百度网盘
• Google Drive

• online hard examples mining
• soft-nms
• TTA (multi-scale testing, flip and box voting)
• multi-scale training

This work builds on many excellent works, which include:

• matterport/Mask_RCNN
• open-mmlab/mmdetection