David Silva

@rose-jinyang, the original paper suggests that there's independence between the number of iterations and the resulting curve. My own experience tells me that it depends on the dataset and the...

That would mean that `custom_func` must have a backward pass defined, otherwise you won't be able to do backpropagation through it. If that's the case, why not just call the...

sadly no, not sure if the algorithm would even provide any meaningful results for things like gans. As far as I know, there's no implementation of this algorithm that supports...

A DataLoader is not strictly needed. It requires an instance of [`TrainDataLoaderIter`](https://github.com/davidtvs/pytorch-lr-finder/blob/acc5e7ee7711a460bf3e1cc5c5f05575ba1e1b4b/torch_lr_finder/lr_finder.py#L51) and optionally [`ValDataLoaderIter`](https://github.com/davidtvs/pytorch-lr-finder/blob/acc5e7ee7711a460bf3e1cc5c5f05575ba1e1b4b/torch_lr_finder/lr_finder.py#L70). You can implement classes that inherit from these and modify their behavior. The important thing...

How exactly are you running? What are the command-line arguments?

The annotations of the Cityscapes test dataset are not public and since you have `--imshow-batch` it tries to display the annotation which then ends up causing the error you got....

@lucky26418 , sorry for the late reply. Hope that in meantime you have resolved the issue since unfortunately, I could not reproduce your error using the command you posted before:...

@Stone-SL, please post here the command-line arguments that you are using and the error trace

@Stone-SL, I'm unable to reproduce your error. If I run the command you posted on my machine the model is loaded and run on the test dataset. The output while...

Hi, I never tried training with full resolution images so I don't really know what is a normal score. I can give it a try but don't really know when...