Regarding training a mask model with my own data, could you please provide guidance on the steps involved about maskgit-pytorch

[llvictorll]

Hello,

If you want to train Maskgit on your custom data, you need to provide a pytorch Dataset for your specific dataset and add it in get_data, with a custom name (i.e. "custom_data"). Next, you just need to run the following cmd (for a single gpu):

data_folder="/path/to/our/new/dataset/"
data="custom_data"
vit_folder="/path/to/save/your/maskgit"
vqgan_folder="/path/to/your/pretrained/vqgan"
writer_log="./logs/"
num_worker=16
bsize=256
python main.py --bsize ${bsize} --data-folder "${data_folder}" --vit-folder "${vit_folder}" --vqgan-folder "${vqgan_folder}" --writer-log "${writer_log}" --num_workers ${num_worker} --data ${data} --img-size 256 --epoch 300