preprocess scripts for LIP

[shouyanxiang]

Thanks for your work!

I download LIP dataset from here, and get dataset folder structure as below:

.
|-- ATR
|   `-- humanparsing
|       |-- JPEGImages
|       `-- SegmentationClassAug
|-- CIHP
|   `-- instance-level_human_parsing
|       |-- Testing
|       |   `-- Images
|       |-- Training
|       |   |-- Categories
|       |   |-- Category_ids
|       |   |-- Human
|       |   |-- Human_ids
|       |   |-- Images
|       |   |-- Instance_ids
|       |   `-- Instances
|       `-- Validation
|           |-- Categories
|           |-- Category_ids
|           |-- Human
|           |-- Human_ids
|           |-- Images
|           |-- Instance_ids
|           `-- Instances
`-- LIP

that is different from the structure you mentioned in GETTING_STARTED.md:

├── lip
│   ├── atr
│   │   ├── edge
│   │   ├── image
│   │   └── label
│   ├── cihp
│   │   ├── image
│   │   └── label
│   ├── train
│   │   ├── edge
│   │   ├── image
│   │   └── label
│   ├── val
│   │   ├── edge
│   │   ├── image
│   │   └── label

could you please provide the scripts to preprocess LIP dataset? Thanks for a lot!