gwinndr
Continue training error and trained model
In train.py line 137
for epoch in range(start_epoch, args.epochs):
When I want to continue to train based on a model weight, there will be an error. Maybe it should be start_epoch+args.epochs ?
I also woud like to ask if you can provide a trained weight, which can be very convenient.
Thank you
what is the error? Also yeah I should probably re-train the model... the weights I got are unfortunately lost.
For example, I first trained 20 epochs and the next time I want to continue to train 20 more epochs based on the last weight "epoch_0020.pickle". So the argument should be -continue_weights ...\epoch_0020.pickle -continue_epoch 20 -epochs 20. So for for epoch in range(start_epoch, args.epochs): the range will be (20, 20), and the script will not run. I do not know if my understanding of -continue_epoch arugment function is correct. Maybe it is my mistake.