ekazakos/grove
Video-Text-to-Text β’ 8B β’ Updated β’ 32
Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
This repo contains the manually annotated dataset, iGround, introduced in the paper "Large-scale Pre-training for Grounded Video Caption Generation".
You can load each configuration and split directly with the π€ Datasets library:
from datasets import load_dataset
repo = "ekazakos/iGround"
# Available configs:
# - data_processed
# - data_raw
# - keys
#
# Each config includes the standard splits: train, val, and test.
# data_processed: annotations after processing used to train GROVE.
# Processing merges multiple instances of the same object type in a clip
# into a single annotation by taking the union of all boxes for that instance.
ds_proc_train = load_dataset(repo, "data_processed", split="train")
ds_proc_val = load_dataset(repo, "data_processed", split="val")
ds_proc_test = load_dataset(repo, "data_processed", split="test")
# data_raw: raw annotations without any processing.
# The same object type can appear multiple times in a video,
# with distinct bounding boxes per instance and per frame.
ds_raw_train = load_dataset(repo, "data_raw", split="train")
ds_raw_val = load_dataset(repo, "data_raw", split="val")
ds_raw_test = load_dataset(repo, "data_raw", split="test")
# keys: contains the corresponding video_ids for the above splits.
ds_keys_train = load_dataset(repo, "keys", split="train")
ds_keys_val = load_dataset(repo, "keys", split="val")
bash scripts/download_iGround.sh iGround_links.txt /path/to/iground_videos_dir
If you use this dataset, please cite:
@inproceedings{kazakos2025grove,
title = {Large-scale Pre-training for Grounded Video Caption Generation},
author = {Evangelos Kazakos and Cordelia Schmid and Josef Sivic},
booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
year = {2025}
}