Research map / Vision-language and multimodal / CLIP and zero-shot
OneLIP: Unlocking and Improving Long-Text Representations of CLIP via One-Stage Adaptation
Renjie Pan, Jiayan Song, Hua Yang
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add TRIM module to CLIP (ablation) | CLIP ViT-B/16 · ShareGPT4V | R@1 I→T: 78.2 -> 93.2 | Improved |
| add TRIM module to CLIP (ablation) | CLIP ViT-B/16 · Urban-1K | R@1 T→I: 53.6 -> 80.0 | Improved |
| add TRIM module to CLIP (ablation) | CLIP ViT-B/16 · Flickr30K | R@1 T→I: 24.7 -> 32.0 | Improved |
| add TRIM module to CLIP (ablation) | CLIP ViT-B/16 · MS-COCO | R@1 I→T: 51.8 -> 55.5 | Improved |
| add TRIM module to CLIP (ablation) | CLIP ViT-B/16 · CIFAR-10 | Acc@1: 90.8 -> 90.4 | Hurt |
| add TRIM module to CLIP (ablation) | CLIP ViT-B/16 · CIFAR-100 | Acc@1: 67.3 -> 68.1 | Improved |
| add offline hard negative mining (M=1024) to TRIM (ablation) | TRIM · ShareGPT4V | R@1 I→T: 93.2 -> 94.3 | Improved |
| add offline hard negative mining (M=1024) to TRIM (ablation) | TRIM · Urban-1K | R@1 T→I: 80.0 -> 83.9 | Improved |
| replace offline HNM (M=1024) with PO-HNM (M=512) (ablation) | TRIM · ShareGPT4V | R@1 I→T: 94.3 -> 96.9 | Improved |
| replace offline HNM (M=1024) with PO-HNM (M=512) (ablation) | TRIM · Urban-1K | R@1 T→I: 83.9 -> 87.0 | Improved |
| double PO-HNM negative queue size from 512 to 1024 (ablation) | TRIM + PO-HNM · ShareGPT4V | R@1 I→T: 96.9 -> 99.1 | Improved |
| double PO-HNM negative queue size from 512 to 1024 (ablation) | TRIM + PO-HNM · Urban-1K | R@1 T→I: 87.0 -> 87.2 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Long-CLIP — Long-CLIP: Unlocking the Long-Text Capability of CLIP (ECCV 2024) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FineLIP: Extending CLIP's Reach via Fine-Grained Alignment with Longer Text Inputs (CVPR 2025 · similarity 0.91)
- Retaining Knowledge and Enhancing Long-Text Representations in CLIP through Dual-Teacher Distillation (CVPR 2025 · similarity 0.91)
- TULIP: Token-length Upgraded CLIP (ICLR 2025 · similarity 0.91)
- TripletCLIP: Improving Compositional Reasoning of CLIP via Synthetic Vision-Language Negatives (NeurIPS 2024 · similarity 0.90)
- SuperCLIP: CLIP with Simple Classification Supervision (NeurIPS 2025 · similarity 0.90)
- Modeling Caption Diversity in Contrastive Vision-Language Pretraining (ICML 2024 · similarity 0.90)
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.
