Trouble finding out command-line arguments

[woctezuma]

I see in the README that there are command-line arguments, such as image-size, and I would like to find out the exhaustive list of possible command-line arguments.

However, I have a bit of trouble understanding the code in stylegan2_pytorch.py. I don't see any main() function. I am familiar with Python, but it seems that you are using it in a way which I don't know about, with that binary file in bin/.

Could you point out where I should look at?

[mulka]

The file in bin/ is a Python file. At the bottom, it looks like it is using the fire package to call the train_from_folder function in that same file. I'm guessing that using "--help" with the command will give you a list of the options, but you could also look at the arguments to the train_from_folder function as well. Let me know if that helps.

[woctezuma]

The file in bin/ is a Python file.

Thank you! This makes things clearer!

I did not know that the file in bin/ was a text file: I assumed it might be some pre-compiled binary file.

[woctezuma]

So, Python Fire is a library to generate command-line interfaces.

The list of command-line arguments, and their default values, are given by the train_from_folder() function, as you mentioned:

def train_from_folder(
    data = './data',
    results_dir = './results',
    models_dir = './models',
    name = 'default',
    new = False,
    load_from = -1,
    image_size = 128,
    network_capacity = 16,
    batch_size = 3,
    gradient_accumulate_every = 5,
    num_train_steps = 100000,
    learning_rate = 2e-4,
    num_workers =  None,
    generate = False,
    num_image_tiles = 8
):
    # [...]