config.yaml

run: acgan_h5_v8
type: train

gpu_id: 1

data_root: "/mnt/data2/bchao/anime/data/dataset.h5"
save_root: "/mnt/data3/bchao/anime/experiments"
runs_root: "/mnt/data3/bchao/anime/logs"

generator_path: "/mnt/data2/bchao/anime/experiments/acgan_h5_v6/ckpt/G_26810.pth"
discriminator_path: "/mnt/data2/bchao/anime/experiments/acgan_h5_v6/ckpt/D_26810.pth"

batch_size: 64
epochs: 5000
print_n_iter: 5
log_n_iter: 5
sample_n_iter: 500
save_n_epoch: 50


optim:
    lr: 0.0002
    beta: 0.5

classes:
    #eye: 10
    #hair: 12
    year: 9

class_num: [2]
select_classes: [[0, 1], [7, 8]]

model:
    noise_dim: 100
    class_dim: 2
    input_size: (64, 64)
    
loss_weight:
    dis: 1
    cls: 1