* [Master branch](https://github.com/ultralytics/yolov3/tree/master): Forward-compatible with all [YOLOv5](https://github.com/ultralytics/yolov5) models and methods (**recommended**).
* [Archive branch](https://github.com/ultralytics/yolov3/tree/archive): Backwards-compatible with original [darknet](https://pjreddie.com/darknet/) *.cfg models (⚠️ no longer maintained).
<imgsrc="https://user-images.githubusercontent.com/26833433/100382066-c8bc5200-301a-11eb-907b-799a0301595e.png"width="1000">** GPU Speed measures end-to-end time per image averaged over 5000 COCO val2017 images using a V100 GPU with batch size 32, and includes image preprocessing, PyTorch FP16 inference, postprocessing and NMS. EfficientDet data from [google/automl](https://github.com/google/automl) at batch size 8.
** AP<sup>test</sup> denotes COCO [test-dev2017](http://cocodataset.org/#upload) server results, all other AP results denote val2017 accuracy.
** All AP numbers are for single-model single-scale without ensemble or TTA. **Reproduce mAP** by `python test.py --data coco.yaml --img 640 --conf 0.001 --iou 0.65`
** Speed<sub>GPU</sub> averaged over 5000 COCO val2017 images using a GCP [n1-standard-16](https://cloud.google.com/compute/docs/machine-types#n1_standard_machine_types) V100 instance, and includes image preprocessing, FP16 inference, postprocessing and NMS. NMS is 1-2ms/img. **Reproduce speed** by `python test.py --data coco.yaml --img 640 --conf 0.25 --iou 0.45`
** All checkpoints are trained to 300 epochs with default settings and hyperparameters (no autoaugmentation).
** Test Time Augmentation ([TTA](https://github.com/ultralytics/yolov5/issues/303)) runs at 3 image sizes. **Reproduce TTA** by `python test.py --data coco.yaml --img 832 --iou 0.65 --augment`
Python 3.8 or later with all [requirements.txt](https://github.com/ultralytics/yolov3/blob/master/requirements.txt) dependencies installed, including `torch>=1.7`. To install run:
YOLOv3 may be run in any of the following up-to-date verified environments (with all dependencies including [CUDA](https://developer.nvidia.com/cuda)/[CUDNN](https://developer.nvidia.com/cudnn), [Python](https://www.python.org/) and [PyTorch](https://pytorch.org/) preinstalled):
- **Google Colab Notebook** with free GPU: <ahref="https://colab.research.google.com/github/ultralytics/yolov3/blob/master/tutorial.ipynb"><imgsrc="https://colab.research.google.com/assets/colab-badge.svg"alt="Open In Colab"></a>
- **Kaggle Notebook** with free GPU: [https://www.kaggle.com/ultralytics/yolov3](https://www.kaggle.com/ultralytics/yolov3)
- **Google Cloud** Deep Learning VM. See [GCP Quickstart Guide](https://github.com/ultralytics/yolov3/wiki/GCP-Quickstart)
- **Docker Image** https://hub.docker.com/r/ultralytics/yolov3. See [Docker Quickstart Guide](https://github.com/ultralytics/yolov3/wiki/Docker-Quickstart) 
detect.py runs inference on a variety of sources, downloading models automatically from the [latest YOLOv3 release](https://github.com/ultralytics/yolov3/releases) and saving results to `runs/detect`.
Download [COCO](https://github.com/ultralytics/yolov3/blob/master/data/scripts/get_coco.sh) and run command below. Training times for YOLOv3/YOLOv3-SPP/YOLOv3-tiny are 6/6/2 days on a single V100 (multi-GPU times faster). Use the largest `--batch-size` your GPU allows (batch sizes shown for 16 GB devices).
Ultralytics is a U.S.-based particle physics and AI startup with over 6 years of expertise supporting government, academic and business clients. We offer a wide range of vision AI services, spanning from simple expert advice up to delivery of fully customized, end-to-end production solutions, including:
For business inquiries and professional support requests please visit us at https://www.ultralytics.com.
## Contact
**Issues should be raised directly in the repository.** For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com.
