Hey there, maybe someone can help me. I am currently using ImageAI to train a project. But I am kinda confused about the outputted models. After 200 Epochs of training I got 2 models, one from epoch 40 and the latest. I kinda assumed that they would be the same model, but as a result from my test, the are not. The docs say
For each increase in the mAP50 after an experiment, a model is saved in the hololens-yolo/models folder. The higher the mAP50, the better the model.
So now I think that the epoch 40 model is the right one and the model latest is the epoch 200 one? And if yes, which model should I really be using to continue training the AI. Thanks in advance
#YoloV3 what model to use after training
6 messages · Page 1 of 1 (latest)
still looking for help
I think Yolov3 might be having a callback - saving the best model and the last model
So model 40 might be better, but check the graphs, if available
Thank you, still don't know how to graph with ImageAI, but will figure it out eventually
Maybe old yolo doesn't produce them