jinpeng0528/CoIN-ASD

CoIN-ASD Benchmark CoIN-ASD is a benchmark dataset designed for multimodal continual instruction tuning (MCIT), based on the CoIN dataset. This dataset aims to evaluate the performance of MCIT models in mitigating essential forgetting. 📝 Paper 🐙 GitHub Dataset Structure The dataset is organized in the following structure: ├── ScienceQA/ │ ├── train_ori.json │ ├── train_x{10,20,40,60,80}.json │ └── test.json ├── TextVQA/ │ ├── train_ori.json │ ├──… See the full description on the dataset page: https://huggingface.co/datasets/jinpeng0528/CoIN-ASD.

<iframe src="https://silytics.si/embed/item/71d5757c-f7d1-4a29-ac07-911dbd1f20d1" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.