Research map / Vision-language and multimodal / CLIP and zero-shot
SAMPLe: A Sharpness Aware Minimization based Optimizer for Prompt Learning in Vision-Language Models
Hossein Rajoli nowdeh, Fatemeh Lotfi, Niloufar Alipour Talemi, Hossein Kashiani, Xiaolong Ma, Fatemeh Afghah
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 SAMPLe optimizer to CoOp | CoOp · Base-to-new, 11-dataset avg | HM: 77.65 -> 78.88 (vs SAGM); 71.66 baseline | Improved |
| add SAMPLe optimizer to CoCoOp | CoCoOp · Base-to-new, 11-dataset avg | HM: 78.28 -> 79.51 (vs SAGM) | Improved |
| add SAMPLe optimizer to MaPLe | MaPLe · Base-to-new, 11-dataset avg | HM: 79.96 -> 80.28 (vs SAGM) | Improved |
| add SAMPLe optimizer to CoPrompt | CoPrompt · Base-to-new, 11-dataset avg | HM: 81.49 -> 82.02 (vs SAGM) | Improved |
| add SAMPLe optimizer to TCP | TCP · Base-to-new, 11-dataset avg | HM: 81.40 -> 81.95 (vs SAGM) | Improved |
| add SAMPLe optimizer to CoOp | CoOp · Cross-dataset zero-shot | avg accuracy: 63.88 -> 65.85 | Improved |
| add SAMPLe optimizer to TCP | TCP · Cross-dataset zero-shot | avg accuracy: 66.29 -> 67.53 | Improved |
| add SAMPLe optimizer to CoOp | CoOp · Domain generalization | avg accuracy: 59.28 -> 60.41 | Improved |
| add SAMPLe optimizer to TCP | TCP · Domain generalization | avg accuracy: 60.51 -> 61.90 | Improved |
| add SAMPLe to CoOp on source ImageNet accuracy | CoOp · Domain generalization (source ImageNet) | accuracy: 71.51 -> 70.60 | Hurt |
| add FSAM to CoOp on ImageNet base-to-new | CoOp · Base-to-new, ImageNet | base accuracy: 76.47 -> 69.79 | Hurt |
| add gradient-alignment term to SAM objective (L + Lp - alpha grad.grad) (ablation) | CoPrompt · Base-to-new, 11-dataset avg | HM: 80.99 -> 81.49 | Improved |
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.
- MaPLe — MaPLe: Multi-Modal Prompt Learning (CVPR 2023) · better on 3 of 3 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
- IPO: Interpretable Prompt Optimization for Vision-Language Models (NeurIPS 2024 · similarity 0.93)
- Learning to Learn Better Visual Prompts (AAAI 2024 · similarity 0.92)
- ZIP: An Efficient Zeroth-order Prompt Tuning for Black-box Vision-Language Models (ICLR 2025 · similarity 0.91)
- SharpZO: Hybrid Sharpness-Aware Vision Language Model Prompt Tuning via Forward-Only Passes (NeurIPS 2025 · similarity 0.91)
- Make Prompts Adaptable: Bayesian Modeling for Vision-Language Prompt Learning with Data-Dependent Prior (AAAI 2024 · similarity 0.91)
- Prompt Learning via Meta-Regularization (CVPR 2024 · similarity 0.91)
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.
