Explanation of parameters in config.json

experiencor / keras-yolo2

Easy training on custom dataset. Various backends (MobileNet and SqueezeNet) supported. A YOLO demo to detect raccoon run entirely in brower is accessible at https://git.io/vF7vI (not on Windows).

MIT License

1.73k stars 787 forks source link

Explanation of parameters in config.json #427

Open yumna4 opened 5 years ago

yumna4 commented 5 years ago

Hi, what do the following parameters mean in the config.json file? Thank you!

"train_times": 8, "nb_epochs": 1, "warmup_epochs": 3, "object_scale": 5.0 , "no_object_scale": 1.0, "coord_scale": 1.0, "class_scale": 1.0,