-
Notifications
You must be signed in to change notification settings - Fork 4
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
2 changed files
with
209 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,104 @@ | ||
# Training | ||
export OPENAI_LOGDIR='OUTPUT/consep_multi-SDM-256CH' \ | ||
# mpiexec -n 8 | ||
python image_train.py \ | ||
--data_dir /Dataset/consep \ | ||
--dataset_mode consep \ | ||
--lr 1e-4 \ | ||
--batch_size 16 \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--use_checkpoint True \ | ||
--num_classes 5 \ | ||
--class_cond True \ | ||
--no_instance False \ | ||
--save_interval 100 \ | ||
--gpu 1 | ||
|
||
# Classifier-free Finetune | ||
export OPENAI_LOGDIR='OUTPUT/consep-SDM-256CH-FINETUNE' | ||
python image_train.py \ | ||
--data_dir /Dataset/consep \ | ||
--dataset_mode consep \ | ||
--lr 2e-5 \ | ||
--batch_size 16 \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--use_checkpoint True \ | ||
--num_classes 5 \ | ||
--class_cond True \ | ||
--drop_rate 0.2 \ | ||
--no_instance False \ | ||
--resume_checkpoint OUTPUT/consep-SDM-256CH-FINETUNE/model005900.pt \ | ||
--save_interval 100 \ | ||
--gpu 3 | ||
|
||
# Testing (orig SDM) | ||
export OPENAI_LOGDIR='OUTPUT/consep-SDM-256CH-TEST' | ||
python image_sample.py \ | ||
--data_dir /Dataset/consep \ | ||
--dataset_mode consep \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--num_classes 5 \ | ||
--class_cond True \ | ||
--batch_size 2 \ | ||
--num_samples 3 \ | ||
--no_instance False \ | ||
--model_path OUTPUT/consep-SDM-256CH-FINETUNE/ema_0.9999_010000.pt \ | ||
--results_path RESULTS/consep-SDM-256CH --s 1.5 \ | ||
--use_ddim \ | ||
--gpu 2 | ||
|
||
# Testing | ||
export OPENAI_LOGDIR='OUTPUT/consep-SDM-256CH-TEST' | ||
python image_sample_repaint.py \ | ||
--data_dir /Dataset/consep \ | ||
--dataset_mode consep \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--num_classes 5 \ | ||
--class_cond True \ | ||
--batch_size 2 \ | ||
--num_samples 3 \ | ||
--no_instance False \ | ||
--model_path OUTPUT/consep-SDM-256CH-FINETUNE/ema_0.9999_010000.pt \ | ||
--results_path RESULTS/consep-SDM-256CH \ | ||
--s 1.5 \ | ||
--gpu 2 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,105 @@ | ||
# Training | ||
export OPENAI_LOGDIR='OUTPUT/glysac-SDM-256CH' | ||
# mpiexec -n 8 | ||
python image_train.py \ | ||
--data_dir /Dataset/glysac \ | ||
--dataset_mode glysac \ | ||
--lr 1e-4 \ | ||
--batch_size 16 \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--use_checkpoint True \ | ||
--num_classes 4 \ | ||
--class_cond True \ | ||
--no_instance False \ | ||
--save_interval 120 \ | ||
--gpu 0 | ||
|
||
# Classifier-free Finetune | ||
export OPENAI_LOGDIR='OUTPUT/glysac-SDM-256CH-FINETUNE' | ||
python image_train.py \ | ||
--data_dir /Dataset/glysac \ | ||
--dataset_mode glysac \ | ||
--lr 2e-5 \ | ||
--batch_size 16 \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--use_checkpoint True \ | ||
--num_classes 4 \ | ||
--class_cond True \ | ||
--drop_rate 0.2 \ | ||
--no_instance False \ | ||
--resume_checkpoint OUTPUT/glysac-SDM-256CH-FINETUNE/model009480.pt \ | ||
--save_interval 120 \ | ||
--gpu 1 | ||
|
||
# Testing (orig SDM) | ||
export OPENAI_LOGDIR='OUTPUT/glysac-SDM-256CH-TEST' | ||
python image_sample.py \ | ||
--data_dir /Dataset/glysac \ | ||
--dataset_mode glysac \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--num_classes 4 \ | ||
--class_cond True \ | ||
--batch_size 2 \ | ||
--num_samples 3 \ | ||
--no_instance False \ | ||
--model_path OUTPUT/glysac-SDM-256CH-FINETUNE/ema_0.9999_010200.pt \ | ||
--results_path RESULTS/glysac-SDM-256CH \ | ||
--s 2.0 \ | ||
--use_ddim \ | ||
--gpu 2 | ||
|
||
# Testing | ||
export OPENAI_LOGDIR='OUTPUT/glysac-SDM-256CH-TEST-fgbg' | ||
python image_sample_repaint.py \ | ||
--data_dir /Dataset/glysac \ | ||
--dataset_mode glysac \ | ||
--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 True \ | ||
--use_scale_shift_norm True \ | ||
--num_classes 4 \ | ||
--class_cond True \ | ||
--batch_size 2 \ | ||
--num_samples 3 \ | ||
--no_instance False \ | ||
--model_path OUTPUT/glysac-SDM-256CH-FINETUNE/ema_0.9999_010200.pt \ | ||
--results_path RESULTS/glysac-SDM-256CH \ | ||
--s 2.0 \ | ||
--gpu 2 |