README.md

December 12, 2023 ยท View on GitHub

Uni-NLX: Unifying Textual Explanations for Vision and Vision-Language Tasks



[arXiv]
[video presentation at ICCV]

Requirements

Images Download

Data

The trianing and test data (combined for all datasets) can be found here

Annotations

The annotations in the format that cococaption expects can be found here. Please place them inside the cococaption folder.

Code

train_nlx.py: script for training only
test_datasets.py: script for validation/testing for all epochs on all 7 NLE tasks
clip_model.py: script for vision backbone we use (CLIP visual encoder)

Models

Results