trained 60 epoch,the validation loss of paf and heatmap is still decrease

[hudengjunai]

hello,I pull down your code almost no change.I trained the mobilenet based model on 2080Ti with 60 epoch.and I found the val_loss is still decrease the x-axis is the epoch from 1 to 60,the y-axis is loss.

[michalfaber]

That's good. The lower val loss, the better.

[fangzhiyuan1995]

Excuse me,Are you using the same learning strategy as train_custom_loop_mobilenet.py?