Questions on running train processes

[tobymuller233]

Hi! I'm recently working on this repo and I'm trying to run some basic experiments mentioned in the paper. For example, applying LeNet on MNIST dataset. In the train_mnist.py, pretrained models are required. In the config directory, there are tons of config files of different names, for example, in config/mnist, there are many yaml files with prefix Lenet,

representing config files for Lenet. However, as I explained, pretrained models are required when training, how can I find pretrained weights for these tons of different configurations? Is there any download link that you can provide? Thanks!

[tobymuller233]

By the way, what do these two methods refer to?

This line of code is in neumeta/train_mnist.py:13

[tobymuller233]

Hi! I'm recently working on this repo and I'm trying to run some basic experiments mentioned in the paper. For example, applying LeNet on MNIST dataset. In the train_mnist.py, pretrained models are required. In the config directory, there are tons of config files of different names, for example, in config/mnist, there are many yaml files with prefix Lenet, representing config files for Lenet. However, as I explained, pretrained models are required when training, how can I find pretrained weights for these tons of different configurations? Is there any download link that you can provide? Thanks!

I found that these config files refer to different configuration for INR training, pretrained weights should be obtained by training it by myself!

[tobymuller233]

By the way, what do these two methods refer to? This line of code is in neumeta/train_mnist.py:13

This issue still bothers me 😭 I can't find the definition of these two methods, neither can vscode.

when I try to run train_mnist.py, an error occurs: