AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

IRGPT: Understanding Real-world Infrared Image with Bi-cross-modal Curriculum on Large-scale Benchmark

Zhe Cao, Jin Zhang, Ruiheng Zhang

ICCV 2025 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment · 8 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add bi-cross-modal curriculum pre-training vs random order (ablation)IRGPT · IR-TD zero-shotpsum: 307.18 -> 328.65Improved
add bi-cross-modal curriculum pre-training vs random order (ablation)IRGPT · IR-TD zero-shotnsum: 97.49 -> 93.87Improved
use easy-to-hard curriculum instead of hard-to-easy (ablation)IRGPT · IR-TD fine-tunedpsum: 433.24 -> 485.79Improved
use curriculum instead of random order (ablation)IRGPT · IR-TD fine-tunednsum: 12.94 -> 4.39Improved
add curriculum-pretrained IRGPT to InternVL2-8B baselineInternVL2-8B · IR-TD zero-shotpsum: 252.30 -> 328.65Improved
add curriculum-pretrained IRGPT to InternVL2-8B baselineInternVL2-8B · IR-TD zero-shotnsum: 125.43 -> 93.87Improved
fine-tuned IRGPT (8B) vs larger InternVL2-26BInternVL2-26B · IR-TD fine-tunedpsum: 473.78 -> 485.79Improved
random pre-training beats curriculum on pedestrian countingIRGPT · IR-TD zero-shotPedestrian Counting MAE: 47.17 vs 47.30Hurt
add loss variation rate weight alpha (ablation)IRGPT (L1+L2) · IR-TD fine-tunedpsum: 481.36 -> 485.79Improved
add loss variation rate weight alpha (ablation)IRGPT (L1+L2) · IR-TD fine-tunednsum: 7.08 -> 4.39Improved
add Lesson 2 (IR-T) to Lesson 1 curriculum (ablation)IRGPT · IR-TD fine-tunedpsum: 466.81 -> 485.79Improved
add Lesson 1 (IR-VIS) to Lesson 2 curriculum (ablation)IRGPT · IR-TD fine-tunedpsum: 470.23 -> 485.79Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy