dataset/README.md

You can download these datasets: COCO, CrowdPose, MPII, AIC, COCO-Stuff, RefCOCO, GrefCOCO, GoPro, REDS, SIDD, CLWD, IP2PDataset, GIER, GQAInpaint, MagicBrush. The resulting data directory should look like this:

InstructDiffusion
|-- data
`-- |-- coco
    |   |-- annotations
    |   `-- images
    |-- mpii
    |   |-- annot
    |   `-- images
    |-- crowdpose
    |   |-- json
    |   `-- images
    |-- aic
    |   |-- annotations
    |   `-- ai_challenger_keypoint_train_20170902
    |
    |-- coco-stuff
    |   |-- annotations
    |   |-- labels.txt
    |   `-- images
    |-- coco_2014
    |   |-- grefcoco
    |   |   |-- grefs(unc).json
    |   |   `-- instances.json
    |   |-- refcoco
    |   |   |-- instances.json
    |   |   |-- refs(google).p
    |   |   `-- refs(unc).p
    |   `-- images
    |        
    |-- GoPro
    |   |-- train
    |   `-- test
    |-- REDS
    |   |-- train
    |   `-- val
    |-- SIDD
    |   |-- train
    |   `-- val
    |-- CLWD
    |   |-- train
    |   |-- test
    |   `-- watermark_logo
    |
    |-- clip-filtered-dataset
    |   |-- shard-00.zip
    |   |-- shard-01.zip
    |   `-- ... 
    |-- GIER_editing_data
    |   |-- images
    |   `-- GIER.json
    |-- gqa-inpaint
    |   |-- images
    |   |-- images_inpainted
    |   |-- masks
    |   |-- train_scenes.json
    |   `-- meta_info.json
    `-- MagicBrush
        |-- data
        |-- processed-train
        `-- magic_train.json