Research map / Federated learning and privacy / Personalized and heterogeneous FL
TOFA: Training-Free One-Shot Federated Adaptation for Vision-Language Models
Li Zhang, Zhongxuan Han, XiaoHua Feng, Jiaming Zhang, Yuyuan Li, Linbo Jiang, Jianan Lin, Chaochao Chen
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 |
|---|---|---|---|
| fuse visual and textual pipelines instead of a single modality (ablation) | TOFA · CLIP datasets, DomainNet, Office-Caltech10 | accuracy | Improved |
| give higher weight to global information in local posterior (alpha >= 0.75) (ablation) | TOFA visual pipeline · CLIP datasets (10 clients) | accuracy | Improved |
| increase number of shots from 1 to 16 (ablation) | TOFA · CLIP datasets (10 clients) | accuracy | Improved |
| use TOFA training-free one-shot adaptation instead of one-shot baselines with PromptFL | federated CLIP adaptation · DomainNet (feature shift) | average accuracy: 82.73 -> 93.05 | Improved |
| use TOFA instead of multi-round prompt learning (DP-PFL) | federated CLIP adaptation · OxfordPets, Flowers102, Food101, DTD | accuracy: OxfordPets 91.23 vs 96.91; DTD 71.68 vs 86.23 | Mixed |
| use TOFA instead of pure training-free methods (Zero-Shot CLIP, CLIP-GDA) | federated CLIP adaptation · DomainNet (feature shift) | average accuracy: 66.71 -> 93.05 (vs CLIP-GDA) | Improved |
Closest papers by meaning
- F^3OCUS - Federated Finetuning of Vision-Language Foundation Models with Optimal Client Layer Updating Strategy via Multi-objective Meta-Heuristics (CVPR 2025 · similarity 0.89)
- Mixture of Experts Made Personalized: Federated Prompt Learning for Vision-Language Models (ICLR 2025 · similarity 0.89)
- BLEND: Balancing Personalization vs. Generalization in Federated Vision–Language Models (NeurIPS 2026 · similarity 0.88)
- Ravan: Multi-Head Low-Rank Adaptation for Federated Fine-Tuning (NeurIPS 2025 · similarity 0.88)
- Harmonizing Generalization and Personalization in Federated Prompt Learning (ICML 2024 · similarity 0.88)
- Global Prompt Refinement with Non-Interfering Attention Masking for One-Shot Federated Learning (NeurIPS 2025 · similarity 0.87)
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.
