Jen-Vu / bottom_try_on

Bottom try on

Data preprocessing

We convert the original data VITON into different directories for easily use.

You can get the processed data at GoogleDrive or by running:

Geometric Matching Module

training

We just use L1 loss for criterion in this code.

TV norm constraints for the offsets will make GMM more robust.

An example training command is

python train.py --name gmm_train_new --stage GMM --workers 4 --save_count 5000 --shuffle

eval

Choose the different source data for eval with the option --datamode.

An example training command is

python test.py --name gmm_traintest_new --stage GMM --workers 4 --datamode test --data_list test_pairs.txt --checkpoint checkpoints/gmm_train_new/gmm_final.pth

You can see the results in tensorboard, as show below.

Try-On Module

background keep

Using cp_dataset.py for keep background and cp_dataset_old for remove background

training

Before the trainning, you should generate warp-mask & warp-cloth, using the test process of GMM with --datamode train. Then move these files or make soft links under the directory data/train. An example training command is

python train.py --name tom_train_new --stage TOM --workers 4 --save_count 5000 --shuffle 

You can see the results in tensorboard, as show below.

eavl

An example training command is

python test.py --name tom_test_new --stage TOM --workers 4 --datamode test --data_list test_pairs.txt --checkpoint checkpoints/tom_train_new/tom_final.pth

You can see the results in tensorboard, as show below.

Pretrain model

You can download here