Research map / 3D vision / Face and shape
Transformers As Meta-Learners for Implicit Neural Representations
Yinbo Chen, Xiaolong Wang
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 |
|---|---|---|---|
| increase weight groups G from 1 to 64 (ablation) | Transformer meta-learner · CelebA image regression | PSNR: 25.63 -> 31.96 | Improved |
| increase weight groups from 1 to 4 (ablation) | Transformer meta-learner · CelebA image regression | PSNR: 25.63 -> 27.89 | Improved |
| increase weight groups from 4 to 16 (ablation) | Transformer meta-learner · CelebA image regression | PSNR: 27.89 -> 29.93 | Improved |
| increase weight groups from 16 to 64 (ablation) | Transformer meta-learner · CelebA image regression | PSNR: 29.93 -> 31.96 | Improved |
| replace gradient-based Learned Init with Transformer hypernetwork | INR image regression · CelebA | PSNR: 30.37 -> 31.96 | Improved |
| replace Learned Init with Transformer hypernetwork | INR image regression · Imagenette | PSNR: 27.07 -> 29.01 | Improved |
| replace Learned Init with Transformer hypernetwork | single-view NeRF · ShapeNet chairs | PSNR: 18.85 -> 19.66 | Improved |
| replace Learned Init with Transformer hypernetwork | single-view NeRF · ShapeNet cars | PSNR: 22.80 -> 23.78 | Improved |
| replace Learned Init with Transformer hypernetwork | single-view NeRF · ShapeNet lamps | PSNR: 22.35 -> 22.76 | Improved |
| add 100 test-time optimization steps (ablation) | Transformer meta-learner · ShapeNet lamps (1-view) | PSNR: 22.76 -> 24.71 | Improved |
| add 100 test-time optimization steps (ablation) | Transformer meta-learner · ShapeNet chairs (2-view) | PSNR: 21.10 -> 23.59 | Improved |
| add a second input view (ablation) | Transformer meta-learner · ShapeNet chairs (no test-time opt.) | PSNR: 19.66 -> 21.10 | Improved |
Closest papers by meaning
- Polynomial Implicit Neural Representations for Large Diverse Datasets (CVPR 2023 · similarity 0.90)
- Learning Transferable Features for Implicit Neural Representations (NeurIPS 2024 · similarity 0.88)
- Implicit Representations via Operator Learning (ICML 2024 · similarity 0.88)
- Hyper-Transforming Latent Diffusion Models (ICML 2025 · similarity 0.88)
- VI^3NR: Variance Informed Initialization for Implicit Neural Representations (CVPR 2025 · similarity 0.87)
- MIRE: Matched Implicit Neural Representations (CVPR 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.
