Skip to content

detection

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History
 
 

detection

README.md

Applying SimViT to Object Detection

Our detection code is developed on top of MMDetection v2.13.0 and PVT.

For details see SimViT: Exploring a Simple Vision Transformer with sliding windows.

If you use this code for a paper please cite:

SimViT

@misc{li2021simvit,
      title={SimViT: Exploring a Simple Vision Transformer with sliding windows}, 
      author={Gang Li and Di Xu and Xing Cheng and Lingyu Si and Changwen Zheng},
      year={2021},
      eprint={2112.13085},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Usage

Install MMDetection v2.13.0.

or

pip install mmdet==2.13.0 --user

Apex (optional):

git clone https://github.com/NVIDIA/apex
cd apex
python setup.py install --cpp_ext --cuda_ext --user

If you would like to disable apex, modify the type of runner as EpochBasedRunner and comment out the following code block in the configuration files:

fp16 = None
optimizer_config = dict(
    type="DistOptimizerHook",
    update_interval=1,
    grad_clip=None,
    coalesce=True,
    bucket_size_mb=-1,
    use_fp16=True,
)

Data preparation

Prepare COCO according to the guidelines in MMDetection v2.13.0.

Results and models

  • SimViT on COCO
Method Backbone Pretrain Lr schd box AP Config Download
RetinaNet SimViT-Small ImageNet-1K 3x 46.3 config log & model
ATSS SimViT-Small ImageNet-1K 3x 49.6 config log & model
GFL SimViT-Small ImageNet-1K 3x 49.9 config log & model

Evaluation

To evaluate PVT-Small + RetinaNet (640x) on COCO val2017 on a single node with 4 gpus run:

dist_test.sh configs/retinanet_capt_small_fpn_3x_mstrin_fp16.py /path/to/checkpoint_file 4 --out results.pkl --eval bbox

This should give

Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.387
Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=1000 ] = 0.593
Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=1000 ] = 0.408
Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=1000 ] = 0.212
Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=1000 ] = 0.416
Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=1000 ] = 0.544
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.545
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=300 ] = 0.545
Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=1000 ] = 0.545
Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=1000 ] = 0.329
Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=1000 ] = 0.583
Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=1000 ] = 0.721

Training

To train PVT-Small + RetinaNet (640x) on COCO train2017 on a single node with 8 gpus for 36 epochs run:

dist_train.sh configs/retinanet_capt_small_fpn_3x_mstrin_fp16.py 8

Demo

python demo.py demo.jpg /path/to/config_file /path/to/checkpoint_file

Calculating FLOPS & Params

python get_flops.py configs/gfl_pvt_v2_b2_fpn_3x_mstrain_fp16.py

This should give

Input shape: (3, 1280, 800)
Flops: 260.65 GFLOPs
Params: 33.11 M