# ATSS-PyTorch

**Repository Path**: zixingtang/ATSS-PyTorch

## Basic Information

- **Project Name**: ATSS-PyTorch
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-05-09
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# ATSS-EfficientDet-PyTorch

ATSS-EfficientDet ([ATSS](https://arxiv.org/pdf/1912.02424.pdf) built on top of [EfficientDet](https://arxiv.org/pdf/1911.09070.pdf)) ```outperforms``` the original EfficientDet. Weights are released.

This repository is folked from [mmdetection](https://github.com/open-mmlab/mmdetection), implemented in [PyTorch](https://pytorch.org/).


## Change log

- 2020/04/17: Release checkpoint of ATSS-EfficientDet-D0 (COCO-val mAP=33.8). Add config file of ATSS-EfficientDet-D1.


## Introduction

There are more applications using Deep Learning in Computer Vision like Image Classification, Object Detection, and Semantic Segmentation. However, Deep Learning models, for example CNN, are too slow. Therefore, more effort is paid in order to make these models running faster.

This work was activated by the [EfficientDet-family networks](https://arxiv.org/pdf/1911.09070.pdf) in Object Detection, published by Google team, which consume small GFLOPs without accuracy compensation. Besides, [ATSS](https://arxiv.org/pdf/1912.02424.pdf) is a sampling method for Anchor-based architectures (e.g., RetinaNet), aiming at boosting the detection accuracy with only single anchor at each location.

This repository is based on [mmdetection](https://github.com/open-mmlab/mmdetection) and tries to:
  - reimplement EfficientDet in PyTorch
  - build ATSS on top of EfficientDet
  - train ATSS-EfficientDet with small versions (D0, D1)

<p align="center">
  <img src="demo/atss_effdet_d0/demo.jpg" height="250" alt="accessibility text">
  <img src="demo/atss_effdet_d0/img1.jpg" height="250" alt="accessibility text">
  <br>
  <em>Demo images predicted by ATSS-EfficientDet-D0</em>
</p>

See more demo images in [demo/atss_effdet_d0](demo/atss_effdet_d0)


## Comparison

|                           Model                          |                                                  Weight                                                  | Params (M) | FLOPs (G) | COCO-val mAP | Original COCO-val mAP |
|:--------------------------------------------------------:|:--------------------------------------------------------------------------------------------------------:|:----------:|:---------:|:------------:|:---------------------:|
| [ATSS-EfficientDet-D0](configs/effdet/atss_effdet_d0.py) | [atss_effdet_d0.pth](https://drive.google.com/file/d/1sPviqiKxEXsLdtj94LmLmCNQKIfTBy7M/view?usp=sharing) |    3.83    |    2.32   |     33.8     |          33.5         |
| [ATSS-EfficientDet-D1](configs/effdet/atss_effdet_d1.py) |                                                                                                          |    6.56    |    5.65   |              |          39.1         |


## Installation

Please refer to the official guildeline of [mmdetection](https://github.com/open-mmlab/mmdetection) in [INSTALL.md](docs/INSTALL.md) for installation and dataset preparation.


## Training and testing

  - To train model:
  ```
  scripts/train_coco.sh
  ```

  - To test model:
  ```
  scripts/test_coco.sh
  ```


## Tips for training on a custom dataset

  - Assume your custom dataset has 10 (foreground) classes (and 1 background class), the ```num_classes``` should be 11:
  ```
	bbox_head=dict(
		type='ATSSEffDetHead',
		num_classes=11,
    ...
	),
  ```

  - Load the released weights as the pretraining:
  ```
  load_from = "work_dirs/atss_effdet_d0.pth"
  ```

  - Learning rate vs. batchsize and number of GPUs:
  ```
  lr_start = (batchsize * GPUs) / 16 * (1e-2)
  lr_end = lr_start / 100
  ```

  - Number of epochs: The original config file takes 300 epochs to finalize a cosine learning-rate schedule. If you want to train a custom dataset, you should set a smaller number of epochs.


## References

* Two papers:
  - [EfficientDet: Scalable and Efficient Object Detection](https://arxiv.org/pdf/1911.09070.pdf)
  - [Bridging the Gap Between Anchor-based and Anchor-free Detection via Adaptive Training Sample Selection](https://arxiv.org/pdf/1912.02424.pdf)

* Three repositories:
  - [mmdetection](https://github.com/open-mmlab/mmdetection)
  - [automl/efficientdet](https://github.com/google/automl/tree/master/efficientdet)
  - [pytorch-image-models](https://github.com/rwightman/pytorch-image-models)


## Citation

```
@article{atssefficientdet,
  title   = {ATSS-EfficientDet: ATSS built on top of EfficientDet},
  author  = {Thuy Nguyen-Chinh},
  journal= {https://github.com/thuyngch/ATSS-EfficientDet-PyTorch},
  year={2020}
}
```