MYMY-young/DelimScaling
[ICLR 2026] Official implementation of "Enhancing Multi-Image Understanding Through Delimiter Token Scaling"
When working with advanced AI models that analyze multiple images or documents at once, sometimes these models struggle to keep information from different sources separate, leading to confusion. This project provides a method to improve how these AI models understand and distinguish between various inputs, ensuring more accurate analysis. It takes your existing multi-image or multi-document AI model and makes it better at handling multiple distinct items.
Use this if you are a machine learning engineer or researcher experiencing issues with large vision-language models or large language models confusing information across multiple images or documents in a single query.
Not ideal if you are working with single-image or single-document AI tasks, or if you need to perform full model retraining rather than an inference-time enhancement.
Stars
14
Forks
—
Language
Python
License
MIT
Category
Last pushed
Feb 25, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/MYMY-young/DelimScaling"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
KimMeen/Time-LLM
[ICLR 2024] Official implementation of " 🦙 Time-LLM: Time Series Forecasting by Reprogramming...
om-ai-lab/VLM-R1
Solve Visual Understanding with Reinforced VLMs
bytedance/SALMONN
SALMONN family: A suite of advanced multi-modal LLMs
NVlabs/OmniVinci
OmniVinci is an omni-modal LLM for joint understanding of vision, audio, and language.
fixie-ai/ultravox
A fast multimodal LLM for real-time voice