Senqiao/

LISA_Plus_Conversations

🤗Data | 📄Paper | 🚀Code | 💻Model | 🔥Citation

Dataset type: The LISA++ Conversation dataset is a QA dataset designed to train MLLM models with the ability to perform segmentation in dialogue. It is built upon the COCO2017 dataset. Where to send questions or comments about the dataset: https://github.com/dvlab-research/LISAPaper: https://arxiv.org/abs/2312.17240 This… See the full description on the dataset page: https://huggingface.co/datasets/Senqiao/LISA_Plus_Conversations.