/TTFNet-detectron2

Unofficial implementation for TTFNet object detection

Primary LanguagePython

Training-Time-Friendly Network for Real-Time Object Detection

Unofficial implementation for TTFNet instance segmentation.

The weights will be published once the training is complete.

image-20210518000132433

Install

The code is based on detectron2. Please check Install.md for installation instructions.

Training

Follows the same way as detectron2.

Single GPU:

python train_net.py --config-file configs/centernet_r_50_C4_1x.yaml

Multi GPU(for example 8):

python train_net.py --num-gpus 8 --config-file configs/centernet_r_50_C4_1x.yaml

Please adjust the IMS_PER_BATCH in the config file according to the GPU memory.

Inference

Single GPU:

python train_net.py --config-file configs/centernet_r_50_C4_1x.yaml --eval-only MODEL.WEIGHTS /path/to/checkpoint_file

Multi GPU(for example 8):

python train_net.py --num-gpus 8 --config-file configs/centernet_r_50_C4_1x.yaml --eval-only MODEL.WEIGHTS /path/to/checkpoint_file