agile_encoder_celebahq1024x1024_lr_1e-4_150k.py

_base_ = [
    '../_base_/models/agile_encoder.py',
    '../_base_/datasets/unconditional_imgs_rgb_flip_256x256.py',
    '../_base_/default_runtime.py'
]

optimizer = None
# define dataset
# you must set `samples_per_gpu`
# `samples_per_gpu` and `imgs_root` need to be set.
imgs_root = 'data/imgs_256'
data = dict(samples_per_gpu=8,
            workers_per_gpu=4,
            train=dict(dataset=dict(imgs_root=imgs_root)))

# adjust running config
lr_config = None
checkpoint_config = dict(interval=1000, by_epoch=False, max_keep_ckpts=10)
custom_hooks = [
    dict(type='MMGenVisualizationHook',
         output_dir='training_samples',
         res_name_list=['orig_image_bgr', 'inversion_image_bgr'],
         interval=200)
]
log_config = dict(interval=100, hooks=[dict(type='TextLoggerHook')])
# 30000 images in celeba-hq
# total_iters = 500000
total_iters = 150000

# use ddp wrapper for faster training
use_ddp_wrapper = True
find_unused_parameters = False

runner = dict(
    type='DynamicIterBasedRunner',
    is_dynamic_ddp=False,  # Note that this flag should be False.
    pass_training_status=True)