how to save and evaluate the model in the im2im_pred folder? about mtan

[Manchery]

Hi @Sunyingchun

For saving the model, it is a basic operation in Pytorch. If you are familiar with torch, you can easily add your code to save and load model with model.state_dict(), torch.save and torch.load.

For evaluating the model, NYUv2 dataset has only two splits, namely train set and validate set. What the code prints in the screen are exactly evaluating the training model on the validate set.

Please forgive my race to be the first to answer the question @lorenmt. Please correct me if any mistake made.

[Manchery]

I think there is a similar issue: model_segnet_mtan prediction code · Issue #31 · lorenmt/mtan

[syc10-09]

Thanks for your quick answer.