Skip to content

Latest commit

 

History

History

maskrcnn

README.md

Mask R-CNN with TransNeXt backbone on COCO

Model Zoo

COCO object detection and instance segmentation results using the Mask R-CNN method:

Backbone Pretrained Model Lr Schd box mAP mask mAP #Params Download Config Log
TransNeXt-Tiny ImageNet-1K 1x 49.9 44.6 47.9M model config log
TransNeXt-Small ImageNet-1K 1x 51.1 45.5 69.3M model config log
TransNeXt-Base ImageNet-1K 1x 51.7 45.9 109.2M model config log
  • When we checked the training logs, we found that the mask mAP and other detailed performance of the Mask R-CNN using the TransNeXt-Tiny backbone were even better than reported in the paper (versions V1 and V2). We have already fixed this in version V3 (it should be a data entry error).

Requirements

pip install -r requirements.txt

Data preparation

cd /path/to/current_folder
ln -s /your/path/to/coco-dataset ./data

Evaluation

To evaluate Mask R-CNN models with TransNeXt backbone on COCO val, you can use the following command:

bash dist_test.sh <config-file> <checkpoint-path> <gpu-num> --eval bbox segm

For example, to evaluate the TransNeXt-Tiny on a single GPU:

bash dist_test.sh ./configs/mask_rcnn_transnext_tiny_fpn_1x_coco.py /path/to/checkpoint_file 1 --eval bbox segm

For example, to evaluate the TransNeXt-Tiny on 8 GPUs:

bash dist_test.sh ./configs/mask_rcnn_transnext_tiny_fpn_1x_coco.py /path/to/checkpoint_file 8 --eval bbox segm

Training

In order to train Mask R-CNN models with TransNeXt backbone on the COCO dataset, first, you need to fill in the path of your downloaded pretrained checkpoint in ./configs/<config-file>. Specifically, change it to:

pretrained=<path-to-checkpoint>,

After setting up, to train TransNeXt on COCO dataset, you can use the following command:

bash dist_train.sh <config-file> <gpu-num>

For example, to train the TransNeXt-Tiny on 8 GPUs, with a total batch-size of 16:

bash dist_train.sh ./configs/mask_rcnn_transnext_tiny_fpn_1x_coco.py 8

Acknowledgement

The released script for Object Detection with TransNeXt is built based on the MMDetection and timm library.

License

This project is released under the Apache 2.0 license. Please see the LICENSE file for more information.

Citation

If you find our work helpful, please consider citing the following bibtex. We would greatly appreciate a star for this project.

@InProceedings{shi2023transnext,
  author    = {Dai Shi},
  title     = {TransNeXt: Robust Foveal Visual Perception for Vision Transformers},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  month     = {June},
  year      = {2024},
  pages     = {17773-17783}
}