alterzero/RBPN-PyTorch

adjusting learning rate

Closed this issue · 2 comments

when training the network using 9 frames or 11 frames, is it necessary to adjust the learning rate?

I believe training on 9 frames or more will be harder.
When I trained 11 frames, it is fine-tuned from F7's model.
So, you can start from fine-tuning model with lower learning rate.

how many frames did you use to get the best experimental results? @alterzero