Skip to content

Latest commit

 

History

History
44 lines (32 loc) · 1.31 KB

dataset.md

File metadata and controls

44 lines (32 loc) · 1.31 KB

Prepare Datasets

We provide instruction for preparing datasets.

COCO

Please download the COCO dataset, unzip it, and make sure it is in the following structure:

coco/
  annotations/
    instances_{train,val}2017.json
  {train,val}2017/
    # image files that are mentioned in the corresponding json

Then, download instances_train2017_base.json, (instances_train2017_base_RN50relabel.json, instances_train2017_base_RN50x4relabel_pre.json) and instances_val2017_basetarget.json from this Google Drive. They are used for region prompting, localizer training and evaluation respectively. Please put then under coco/annotations.

Export the dataset path (path/to/coco) by executing:

export data_path='path/to/coco'

LVIS

TBD.