| | --- |
| | license: cc-by-nc-nd-4.0 |
| | --- |
| | ## Description |
| | 120,000 Questions - Multi-discipline Multi-modal Understanding & Reasoning Data, covering six major academic fields such as art, engineering, and medicine, as well as various specialized sub-domains. It includes professional-level mixed text and visual questions (e.g., charts, engineering drawings, and art analysis) that have been quality-checked by specialists. This dataset enables deep understanding of textual and visual information and can enhance the logical reasoning and knowledge application capabilities of large models. |
| |
|
| | For more details, please refer to the link: https://www.nexdata.ai/datasets/llm/1751?source=Huggingface |
| |
|
| | ## Specifications |
| | ### Data size |
| |
|
| | 120,000 questions |
| |
|
| | ### Image resolution |
| |
|
| | short side resolution ≥ 500 pixels |
| |
|
| | ### Subject categories |
| |
|
| | arts, business, science, medicine,humanities and social sciences, engineering |
| |
|
| | ### QA length |
| |
|
| | question length ≥ 10 Chinese characters, and the answer and analysis length ≥ 40 characters |
| |
|
| | ### Collection equipment |
| |
|
| | mobile phone, scanner |
| |
|
| | ### Language |
| |
|
| | Chinese |
| |
|
| | ### Diversity |
| |
|
| | multiple disciplines, multiple image types, multiple types of questions |
| |
|
| | ### Data format |
| |
|
| | .jpg, .png, .josn |
| |
|
| | ## Licensing Information |
| | Commercial License |
