No official code found. Best implementation: WenkeHuang/SPIDER (3 stars, last commit 2025-09-24).
Adapt Without Forgetting: Distill Proximity from Dual Teachers in Vision-Language Models
High-Resolution and Few-Shot View Synthesis from Asymmetric Dual-Lens Inputs
Learning from Both Structural and Textual Knowledge for Inductive Knowledge Graph Completion
Absolute Pose from One or Two Scaled and Oriented Features
Sketch2Diagram: Generating Vector Diagrams from Hand-Drawn Sketches
Evolving Minds: Logic-Informed Inference from Temporal Action Patterns
@article{huang2026learn,
title = {Learn from Downstream and Be Yourself in Multimodal Large Language Models Fine-Tuning},
author = {Wenke Huang and Jian Liang and Zekun Shi and Didi Zhu and Guancheng Wan and He Li and Bo Du and Dacheng Tao and Mang Ye},
year = {2026},
url = {https://api.semanticscholar.org/CorpusID:1a7a5de1c49692c2f6d7190840c938a440d7b8b2},
journal = {ICML 2025 2025}
}Shows whether official code is linked for this paper.
[](https://codeofpaper.com/paper/1a7a5de1c49692c2f6d7190840c938a440d7b8b2)