random canny dropout for vid

[wr0124]

random canny for each image in one sequence, and with canny dropout for each image.
The dropout is hard coded for the moment, in "models/palette_model.py"
cond_image_nocanny, frame_select = fill_mask_with_canny_dropout(
self.gt_image,
self.mask,
self.opt.data_temporal_number_frames,
0.9,
)
I print frame_select is to check weather the selected frame in one sequence was suppose to have canny.

python3 -W ignore::UserWarning  train.py \
--dataroot /path/to/online_mario2sonic_full_mario  \
--checkpoints_dir  /path/to/checkpoints \
--name mario_vid_canny_random_dropout0.9_bs1it16_fix1  \
--gpu_ids 0,1,2    \
--model_type palette \
--output_print_freq 100   \
--output_display_freq 500  \
--data_dataset_mode  self_supervised_temporal_labeled_mask_online  \
--train_batch_size 1  \
--train_iter_size 16  \
--model_input_nc 3 \
--model_output_nc 3 \
--data_relative_paths \
--train_G_ema \
--train_optim adamw \
--G_netG unet_vid   \
--data_online_creation_crop_size_A 64  \
--data_online_creation_crop_size_B 64 \
--data_crop_size 64 \
--data_load_size 64  \
--data_online_creation_rand_mask_A \
--train_G_lr 0.0001 \
--dataaug_no_rotate \
--G_diff_n_timestep_train  2000   \
--G_diff_n_timestep_test  1000   \
--data_temporal_number_frames 8  \
--data_temporal_frame_step 1 \
--data_online_creation_mask_delta_A_ratio 0.12 0.12 \
--alg_diffusion_cond_image_creation    computed_sketch  \
--alg_diffusion_cond_computed_sketch_list canny \

[wr0124]

python3 -W ignore::UserWarning train.py
--dataroot /data1/juliew/dataset/online_mario2sonic_full_mario
--checkpoints_dir /data1/juliew/checkpoints
--name mario_vid_canny_random_dropout
--gpu_ids 2
--model_type palette
--output_print_freq 1
--output_display_freq 1
--data_dataset_mode self_supervised_temporal_labeled_mask_online
--train_batch_size 2
--train_iter_size 1
--model_input_nc 3
--model_output_nc 3
--data_relative_paths
--train_G_ema
--train_optim adamw
--G_netG unet_vid
--data_online_creation_crop_size_A 32
--data_online_creation_crop_size_B 32
--data_crop_size 32
--data_load_size 32
--data_online_creation_rand_mask_A
--train_G_lr 0.0001
--dataaug_no_rotate
--G_diff_n_timestep_train 8
--G_diff_n_timestep_test 4
--data_temporal_number_frames 8
--data_temporal_frame_step 1
--data_online_creation_mask_delta_A_ratio 0.12 0.12
--alg_diffusion_cond_image_creation computed_sketch
--alg_diffusion_cond_computed_sketch_list canny
--alg_diffusion_vid_canny_dropout 0.5 \