early stopping in training

[vishnukv64]

Hey,
I just wanted to know how to automatically stop the training if the loss doesnt decrease for say 10 epochs and save the best and last weights?

if there is snothing i could do with the results.txt or something

Thanks

[NanoCode012]

Hm, if you run it without --nosave, you can just cntrl+c the run. Then call ,

yolov5/utils/utils.py

Lines 627 to 633 in 5e2429e

	def strip_optimizer(f='weights/best.pt'): # from utils.utils import *; strip_optimizer()
	# Strip optimizer from *.pt files for lighter files (reduced by 1/2 size)
	x = torch.load(f, map_location=torch.device('cpu'))
	x['optimizer'] = None
	x['model'].half() # to FP16
	torch.save(x, f)
	print('Optimizer stripped from %s' % f)

Put below in yolov5 folder, then call via python <filename>

from utils.utils import strip_optimizer;

strip_optimizer('weights/best.pt')

[glenn-jocher]

https://pytorch.org/docs/stable/optim.html#torch.optim.lr_scheduler.ReduceLROnPlateau might also be of interest.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.