[Website] [Paper] [Code]
Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models. (ICML 2024)
We host VLGuard dataset here. train.json and test.json are the meta data of VLGuard and the images are in train.zip and test.zip.
Please refer to Github for detailed usage.
