GigaGAN-PyTorch

To train a simple unconditional GAN, initialize the GigaGAN class with generator and discriminator configuration dictionaries. You must also use ImageDataset to load your data and call gan.set_dataloader(dataloader) before starting the training loop with gan(steps=..., grad_accum_every=...).

import torch

from gigagan_pytorch import (
    GigaGAN,
    ImageDataset
)

gan = GigaGAN(
    generator = dict(
        dim_capacity = 8,
        style_network = dict(
            dim = 64,
            depth = 4
        ),
        image_size = 256,
        dim_max = 512,
        num_skip_layers_excite = 4,
        unconditional = True
    ),
    discriminator = dict(
        dim_capacity = 16,
        dim_max = 512,
        image_size = 256,
        num_skip_layers_excite = 4,
        unconditional = True
    ),
    amp = True
).cuda()

# dataset

dataset = ImageDataset(
    folder = '/path/to/your/data',
    image_size = 256
)

dataloader = dataset.get_dataloader(batch_size = 1)

# you must then set the dataloader for the GAN before training

gan.set_dataloader(dataloader)

# training the discriminator and generator alternating
# for 100 steps in this example, batch size 1, gradient accumulated 8 times

gan(
    steps = 100,
    grad_accum_every = 8
)

# after much training

images = gan.generate(batch_size = 4) # (4, 3, 256, 256)

To train a Unet Upsampler, set train_upsampler = True in the GigaGAN constructor. The generator configuration should include input_image_size (the resolution of the low-res input). After training, you can generate high-resolution images by passing a low-resolution tensor to gan.generate(lowres).

import torch
from gigagan_pytorch import (
    GigaGAN,
    ImageDataset
)

gan = GigaGAN(
    train_upsampler = True,     # set this to True
    generator = dict(
        style_network = dict(
            dim = 64,
            depth = 4
        ),
        dim = 32,
        image_size = 256,
        input_image_size = 64,
        unconditional = True
    ),
    discriminator = dict(
        dim_capacity = 16,
        dim_max = 512,
        image_size = 256,
        num_skip_layers_excite = 4,
        multiscale_input_resolutions = (128,),
        unconditional = True
    ),
    amp = True
).cuda()

dataset = ImageDataset(
    folder = '/path/to/your/data',
    image_size = 256
)

dataloader = dataset.get_dataloader(batch_size = 1)
gan.set_dataloader(dataloader)

# training the discriminator and generator alternating
# for 100 steps in this example, batch size 1, gradient accumulated 8 times

gan(
    steps = 100,
    grad_accum_every = 8
)

# after much training

lowres = torch.randn(1, 3, 64, 64).cuda()

images = gan.generate(lowres) # (1, 3, 256, 256)