Matteo Guarrera

I would love to have the possibility to use the library directly with grayscale images, without converting them again to RGB

Use the discriminator of the model after training

7

### Problem Description It seems that the library does not have the possibility to use the discriminator after the training. Is it true? ### Expected behavior I believe that it...

If you encounter this error: ``` CMake Error at extern/glfw/CMakeLists.txt:210 (message): The RandR headers were not found -- Configuring incomplete, errors occurred! See also "/home/matteogu/Documents/RoboGrammar/build/CMakeFiles/CMakeOutput.log". See also "/home/matteogu/Documents/RoboGrammar/build/CMakeFiles/CMakeError.log". ``` Use...

Hi there, I am working on long context model. Is it possible to have the pretrained models?

Compatibility Bugs

1

kornia library has refactored some functions and the example/cont_ddpm.py was not working anymore due to an error raised in example/unet.py that needs to be changed to `return kornia.filters.filter2d(x, self.f, normalized=True)`.

bug in the trainer evaluation

1

When passing `test_loaders` the results of the evaluations were overwritten.

Imports in the examples broken

4

`from neuralop.data.datasets import load_darcy_flow_small` this should be `from neuralop.datasets import load_darcy_flow_small` all the examples have this issue :cry:

config.yaml sorted

2

Hi! I think a nice upgrade would be having the configuration yaml file sorted, so it easy to compare different experiments