imagesInternVL-U/ScaleEdit-12Mimage-editingmultimodalinstruction-tuningcomputer-visionimage-to-imagefine-tuningsynthetic-dataparquet

ScaleEdit-12M: Large-Scale Instruction-Based Image Editing Dataset

Name: ScaleEdit-12M: Large-Scale Instruction-Based Image Editing Dataset
Creator: DataBazaar
Keywords: InternVL-U/ScaleEdit-12M, image-editing, multimodal, instruction-tuning, computer-vision, image-to-image, fine-tuning, synthetic-data, parquet

About this data

12.4M verified instruction–image pairs across 23 task families for instruction-based image editing. Largest open-source dataset of its kind, built via the ScaleEditor multi-agent framework. MIT licensed.

Schema

Name	Type	Description
id	BIGINT	Unique sequential identifier for each instruction-image pair in the dataset.
edit_task	VARCHAR	One of 23 editing task families (e.g., count_change, object_removal, style_transfer, attribute_change).
edit_instruction	VARCHAR	Natural-language instruction describing the specific image editing operation to perform.
source_image	BLOB	Input image as JPEG/PNG binary data before editing.
edited_image	BLOB	Output image as JPEG/PNG binary data after applying the edit instruction.
source_image_width	INTEGER	Pixel width of the source image.
source_image_height	INTEGER	Pixel height of the source image.
edited_image_width	INTEGER	Pixel width of the edited image.
edited_image_height	INTEGER	Pixel height of the edited image.
instruction_following_score	INTEGER	Quality score (0–100) measuring how well the edit follows the given instruction.
editing_consistency_score	INTEGER	Quality score (0–100) measuring visual consistency between source and edited images.
generation_quality_score	INTEGER	Quality score (0–100) measuring overall aesthetic and technical quality of the edited image.

Sample Data

Preview a sample of the data before downloading.

Free

Open dataset

Quality: No ratings

0 downloads

Seller: DataBazaar

Agent? No sign-up needed →

For AI Agents

Via MCP Server

# 1. Add to your agent's MCP config (claude_desktop_config.json or similar):
{
  "mcpServers": {
    "databazaar": { "command": "npx", "args": ["databazaar-mcp"] }
  }
}

# 2. Your agent can then call:
search_datasets({ query: "ScaleEdit-12M: Large-Scale Ins" })
// Found: adb934ff-506b-4310-9246-5763d01219c0
get_download_url({ dataset_id: "adb934ff-506b-4310-9246-5763d01219c0" })  // free — no API key needed

Via REST API

# Free dataset — no API key required:
curl https://api.databazaar.io/datasets/adb934ff-506b-4310-9246-5763d01219c0/download-url