which model you choose to finetune?

lliuz / ARFlow

The official PyTorch implementation of the paper "Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimation".

MIT License

251 stars 50 forks source link

which model you choose to finetune? #29

Open RokiaAbdeen opened 3 years ago

RokiaAbdeen commented 3 years ago

you saved two model the last and best modet, which one you used for finetuning? for testing?