We provide instruction tuning files for the 12 datasets included in the CVIT benchmark (including newly added test sets for COCO_Classification and LLaVA_Med), which you can download from the corresponding folders.
All datasets used in the benchmark are publicly available. You can download the corresponding images directly from each dataset’s official website.
The SMoLoRA-specific annotations, data organization, and newly created test files are released under the CC BY 4.0 license.
Some files in this repository are adapted from public datasets and open-source projects, including CoIN. These materials remain subject to their original licenses and terms of use.
Users should also cite SMoLoRA, CoIN, and the corresponding original datasets when using this benchmark.
3 commits
We provide instruction tuning files for the 12 datasets included in the CVIT benchmark (including newly added test sets for COCO_Classification and LLaVA_Med), which you can download from the corresponding folders.
All datasets used in the benchmark are publicly available. You can download the corresponding images directly from each dataset’s official website.
The SMoLoRA-specific annotations, data organization, and newly created test files are released under the CC BY 4.0 license.
Some files in this repository are adapted from public datasets and open-source projects, including CoIN. These materials remain subject to their original licenses and terms of use.
Users should also cite SMoLoRA, CoIN, and the corresponding original datasets when using this benchmark.
3 commits