docs/en/yolov5/tutorials/tips-for-best-training-results.md
š This guide explains how to produce the best mAP and training results with YOLOv5 š.
Most of the time good results can be obtained with no changes to the models or training settings, provided your dataset is sufficiently large and well labeled. If at first you don't get good results, there are steps you might be able to take to improve, but we always recommend users first train with all default settings before considering any changes. This helps establish a performance baseline and spot areas for improvement.
If you have questions about your training results we recommend you provide the maximum amount of information possible if you expect a helpful response, including results plots (train losses, val losses, P, R, mAP), PR curve, confusion matrix, training mosaics, test results and dataset statistics images such as labels.png. All of these are located in your project/name directory, typically yolov5/runs/train/exp.
We've put together a full guide for users looking to get the best results on their YOLOv5 trainings below.
train_batch*.jpg on train start to verify your labels appear correct, i.e. see example mosaic.<a href="https://arxiv.org/abs/1405.0312"></a>
Larger models like YOLOv5x and YOLOv5x6 will produce better results in nearly all cases, but have more parameters, require more CUDA memory to train, and are slower to run. For mobile deployments we recommend YOLOv5s/m, for cloud deployments we recommend YOLOv5l/x. See our README table for a full comparison of all models.
<p align="center"></p>Start from Pretrained weights. Recommended for small to medium-sized datasets (i.e. VOC, VisDrone, GlobalWheat). Pass the name of the model to the --weights argument. Models download automatically from the latest YOLOv5 release.
python train.py --data custom.yaml --weights yolov5s.pt
python train.py --data custom.yaml --weights yolov5m.pt
python train.py --data custom.yaml --weights yolov5l.pt
python train.py --data custom.yaml --weights yolov5x.pt
python train.py --data custom.yaml --weights custom_pretrained.pt
Start from Scratch. Recommended for large datasets (i.e. COCO, Objects365, OIv6). Pass the model architecture YAML you are interested in, along with an empty --weights '' argument:
python train.py --data custom.yaml --weights '' --cfg yolov5s.yaml
python train.py --data custom.yaml --weights '' --cfg yolov5m.yaml
python train.py --data custom.yaml --weights '' --cfg yolov5l.yaml
python train.py --data custom.yaml --weights '' --cfg yolov5x.yaml
Before modifying anything, first train with default settings to establish a performance baseline. A full list of train.py settings can be found in the train.py argparser.
--img 640, though due to the high amount of small objects in the dataset it can benefit from training at higher resolutions such as --img 1280. If there are many small objects then custom datasets will benefit from training at native or higher resolution. Best inference results are obtained at the same --img as the training was run at, i.e. if you train at --img 1280 you should also test and detect at --img 1280.--batch-size that your hardware allows for. Small batch sizes produce poor batch normalization statistics and should be avoided. You can use --batch-size -1 to automatically select the optimal batch size for your GPU.--cos-lr flag to enable cosine learning rate scheduling, which gradually reduces the learning rate following a cosine curve over epochs.mosaic hyperparameter in your --hyp file to help stabilize training.hyp['obj'] will help reduce overfitting in those specific loss components. For an automated method of optimizing these hyperparameters, see our Hyperparameter Evolution Tutorial.--device 0,1,2,3 to distribute training across them, which can significantly reduce training time.--patience 50 to stop training if validation metrics don't improve for 50 epochs, saving time and preventing overfitting.--augment to improve prediction accuracy by averaging results from augmented versions of the input image.If you'd like to know more, a good place to start is Karpathy's 'Recipe for Training Neural Networks', which has great ideas for training that apply broadly across all ML domains: https://karpathy.github.io/2019/04/25/recipe/
For more detailed information on training settings and configurations, refer to the Ultralytics train settings documentation, which provides comprehensive explanations of all available parameters.
Good luck š and let us know if you have any other questions!
Your model may be overfitting if the training loss continues to decrease while validation loss starts to increase. Monitor the validation mAP - if it plateaus or decreases while training loss keeps improving, that's a sign of overfitting. Solutions include adding more training data, increasing data augmentation, or implementing regularization techniques.
The optimal batch size depends on your GPU memory. Larger batch sizes generally provide better batch normalization statistics and training stability. Use the largest batch size your hardware can handle without running out of memory. You can use --batch-size -1 to automatically determine the optimal batch size for your setup.
To speed up training, try: using multiple GPUs with --device 0,1,2,3, caching your dataset with --cache, and optimizing your batch size (mixed precision is enabled automatically on supported GPUs). Also consider using a smaller model variant like YOLOv5s if absolute accuracy isn't critical.