Open minuenergy opened 1 year ago
I want to train photo to photo2 translation(without segmentation mask), How can I train?
I use this train script below
python image_train.py --data_dir data --dataset_mode ade20k --lr 1e-4 --batch_size 1 --attention_resolutions 32,16,8 --diffusion_steps 1000 --image_size 256 --learn_sigma True \ --noise_schedule linear --num_channels 256 --num_head_channels 64 --num_res_blocks 2 --resblock_updown True --use_fp16 False --use_scale_shift_norm True --use_checkpoint True --num_classes 1 \ --class_cond True--no_instance True
and encounter this errors
Hi, did you try it? How did you set the number of classes and this condition? --class_cond True
--class_cond True
I want to train photo to photo2 translation(without segmentation mask), How can I train?
I use this train script below
python image_train.py --data_dir data --dataset_mode ade20k --lr 1e-4 --batch_size 1 --attention_resolutions 32,16,8 --diffusion_steps 1000 --image_size 256 --learn_sigma True \ --noise_schedule linear --num_channels 256 --num_head_channels 64 --num_res_blocks 2 --resblock_updown True --use_fp16 False --use_scale_shift_norm True --use_checkpoint True --num_classes 1 \ --class_cond True--no_instance True
and encounter this errors