Images per class. ≥ 1500 images per class recommended(每个类的图像。≥ 建议每类1500张图片)
Instances per class. ≥ 10000 instances (labeled objects) per class recommended(每个类的实例。≥ 建议每个类10000个实例(标记的对象))
Image variety. Must be representative of deployed environment. For real-world use cases we recommend images from different times of day, different seasons, different weather, different lighting, different angles, different sources (scraped online, collected locally, different cameras) etc.(图像多样性。必须代表已部署的环境。对于真实世界的使用案例,我们推荐来自不同时间、不同季节、不同天气、不同照明、不同角度、不同来源(在线收集、本地收集、不同相机)的图像。)
Label consistency. All instances of all classes in all images must be labelled. Partial labelling will not work.(标签一致性。必须标记所有图像中所有类的所有实例。部分标签将不起作用。)
Label accuracy. Labels must closely enclose each object. No space should exist between an object and it’s bounding box. No objects should be missing a label(标签准确性。标签必须紧紧围绕每个对象。对象与其边界框之间不应存在空间。任何对象都不应缺少标签。)
Label verification. View train_batch*.jpg on train start to verify your labels appear correct, i.e. see example mosaic…(标签验证。在火车开始时查看train_batch*.jpg,以验证标签是否正确,即参见示例马赛克。)
Background images. Background images are images with no objects that are added to a dataset to reduce False Positives (FP). We recommend about 0-10% background images to help reduce FPs (COCO has 1000 background images for reference, 1% of the total). No labels are required for background images.(背景图像。背景图像是没有添加到数据集以减少误报(FP)的对象的图像。我们建议大约0-10%的背景图像,以帮助减少FP(COCO有1000张背景图像供参考,占总数的1%)。背景图像不需要标签。)
Epochs. Start with 300 epochs. If this overfits early then you can reduce epochs. If overfitting does not occur after 300 epochs, train longer, i.e. 600, 1200 etc epochs.
Image size. COCO trains at native resolution of --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. Use the largest --batch-size that your hardware allows for. Small batch sizes produce poor batchnorm statistics and should be avoided.
Hyperparameters. Default hyperparameters are in hyp.scratch-low.yaml. We recommend you train with default hyperparameters first before thinking of modifying any. In general, increasing augmentation hyperparameters will reduce and delay overfitting, allowing for longer trainings and higher final mAP. Reduction in loss component gain hyperparameters like hyp[‘obj’] will help reduce overfitting in those specific loss components. For an automated method of optimizing these hyperparameters, see our Hyperparameter Evolution Tutorial.