DIM-Edit: 100K+ Image Editing Pairs for Unified Multimodal Models

DIM-Edit contains between 100,000 and 1,000,000 records designed to improve precise image editing in unified multimodal models. Released by stdKonjac in October 2025, the data supports the Draw-In-Mind (DIM) framework which rebalances designer and painter roles in diffusion-based architectures. The collection is provided in Parquet format and is associated with the DIM-4.6B model series and Arxiv paper 2509.01986.

Use Cases

• Fine-tuning diffusion models for precise image editing using text-based instructions
• Training unified multimodal models to improve text-to-image alignment
• Researching designer-painter role separation in generative AI architectures

Strengths

• Large scale of 100,000 to 1,000,000 records
• Directly linked to Arxiv 2509.01986 research findings
• CC BY-NC 4.0 license provides clear research usage terms

Limitations

• Non-commercial license restriction
• Lack of detailed column-level documentation in the initial release metadata

Provenance

Source

stdKonjac, Arxiv 2509.01986

Freshness

Released in October 2025; reflects current state-of-the-art research in unified multimodal editing.