Training FFG Models

1. Download dataset

The dataset can be downloaded from here.

2. Prepare decomposition information

(not required for FUNIT)

The files that we used are provided.
- Chinese: data/chn/decomposition.json, data/chn/primals.json
- Korean: data/kor/decomposition.json, data/kor/primals.json, data/kor/decomposition_DM.json(only for DM-Font)
See 3.1 to build your own decomposition file:

3.1 The format of decomposition rule and primals file
- Decomposition rule
  - structure: dict (in json format)
  - format: {char: [list of components]}
  - example: {'㐬': ['亠', '厶', '川'], '㐭': ['亠', '囗', '口']}
- Primals
  - structure: list (in json format)
  - format: [All the components in the decomposition rule file]
  - example: ['亠', '厶', '川', '囗', '口']

3. Modify the data configuration file ("cfgs/data/train/chn_ftransgan.yaml")

1. copy `cfgs/data/train/chn_ftransgan.yaml` to your own configuration file.

Use this command:

cp -f cfgs/data/train/chn_ftransgan.yaml cfgs/data/train/custom.yaml

2. Modify the copied file ("cfgs/data/train/custom.yaml")

Change the (FTransGAN_root) in the copied file to your own dataset root.

Please do not modify the indentation, because the indentation rule is very important in these configuration files.

Evaluating

We provide weights of the classifiers trained with FTransGAN Dataset. (Download)

The list of labels are also provided in "data/chn/ftransgan/eval_keys.json" and "data/chn/ftransgan/eval_chars.json".

1. Modify the data configuration file ("cfgs/evaluator/eval_ftransgan.yaml")

Change the dset.test.data_dir to the root path of your generated images.
Change the (FTransGAN_root) in the copied file to your own dataset root.