| language: | |
| - en | |
| license: apache-2.0 | |
| size_categories: | |
| - 100K<n<1M | |
| task_categories: | |
| - image-to-image | |
| - text-to-image | |
| # Dataset Card for InstructCLIP-InstructPix2Pix-Data | |
| The dataset can be used to train instruction-guided image editing models. | |
| It is built on top of [InstructPix2Pix CLIP-filtered](https://huggingface.co/datasets/timbrooks/instructpix2pix-clip-filtered) with new edit instructions. | |
| For each sample, `source_image` is the original image, `instruction` is the edit instruction, `target_image` is the edited image, and `original_instruction` is the edit instruction from the InstructPix2Pix CLIP-filtered dataset. | |
| Please refer to our [repo](https://github.com/SherryXTChen/Instruct-CLIP.git) to see how the edit instructions are generated, and how to use this dataset to train image editing models. | |
| The data is based on the paper [Instruct-CLIP: Improving Instruction-Guided Image Editing with Automated Data Refinement Using Contrastive Learning](https://huggingface.co/papers/2503.18406). |