PyTorch-YOLOv3

Minimal implementation of YOLOv3 in PyTorch.

Table of Contents

• PyTorch-YOLOv3

• Table of Contents

• Paper

• Installation

• Inference

• Test

• Train

• Credit

Paper

YOLOv3: An Incremental Improvement

_Joseph Redmon, Ali Farhadi_

Abstract
We present some updates to YOLO! We made a bunch of little design changes to make it better. We also trained this new network thats pretty swell. Its a little bigger than last time but more accurate. Its still fast though, dont worry. At 320 320 YOLOv3 runs in 22 ms at 28.2 mAP, as accurate as SSD but three times faster. When we look at the old .5 IOU mAP detection metric YOLOv3 is quite good. It achieves 57.9 AP50 in 51 ms on a Titan X, compared to 57.5 AP50 in 198 ms by RetinaNet, similar performance but 3.8 faster. As always, all the code is online at https://pjreddie.com/yolo/.

[Paper] [Original Implementation]

Installation

$ git clone https://github.com/eriklindernoren/PyTorch-YOLOv3
$ cd PyTorch-YOLOv3/
$ sudo pip3 install -r requirements.txt

Download pretrained weights

$ cd weights/
$ bash download_weights.sh

Download COCO

$ cd data/
$ bash get_coco_dataset.sh

Inference

Uses pretrained weights to make predictions on images. Below table displays the inference times when using as inputs images scaled to 256x256. The ResNet backbone measurements are taken from the YOLOv3 paper. The Darknet-53 measurement marked shows the inference time of this implementation on my 1080ti card.

| Backbone | GPU | FPS | | ----------------------- |:--------:|:--------:| | ResNet-101 | Titan X | 53 | | ResNet-152 | Titan X | 37 | | Darknet-53 (paper) | Titan X | 76 | | Darknet-53 (this impl.) | 1080ti | 74 |

$ python3 detect.py --image_folder /data/samples

Test

Evaluates the model on COCO test.

$ python3 test.py --weights_path weights/yolov3.weights

| Model | mAP (min. 50 IoU) | | ----------------------- |:----------------:| | YOLOv3 (paper) | 57.9 | | YOLOv3 (this impl.) | 58.2 |

Train

Model does not converge yet during training. Data augmentation as well as additional training tricks remains to be implemented. PRs are welcomed!

    train.py [-h] [--epochs EPOCHS] [--image_folder IMAGE_FOLDER]
                [--batch_size BATCH_SIZE]
                [--model_config_path MODEL_CONFIG_PATH]
                [--data_config_path DATA_CONFIG_PATH]
                [--weights_path WEIGHTS_PATH] [--class_path CLASS_PATH]
                [--conf_thres CONF_THRES] [--nms_thres NMS_THRES]
                [--n_cpu N_CPU] [--img_size IMG_SIZE]
                [--checkpoint_interval CHECKPOINT_INTERVAL]
                [--checkpoint_dir CHECKPOINT_DIR]

Credit

@article{yolov3,
  title={YOLOv3: An Incremental Improvement},
  author={Redmon, Joseph and Farhadi, Ali},
  journal = {arXiv},
  year={2018}
}