| Name: | RENet |
|---|---|
| Embed.: | ResNet12 |
| Type: | Fine-tuning |
| Venue: | ICCV'21 |
| Codes: | renet |
Cite this work with:
@article{DBLP:journals/corr/abs-2108-09666,
author = {Dahyun Kang and
Heeseung Kwon and
Juhong Min and
Minsu Cho},
title = {Relational Embedding for Few-Shot Classification},
journal = {CoRR},
volume = {abs/2108.09666},
year = {2021},
}
Classification
| Embedding | 📖 miniImageNet (5,1) | 💻 miniImageNet (5,1) | 📖miniImageNet (5,5) | 💻 miniImageNet (5,5) | 📝 Comments | |
|---|---|---|---|---|---|---|
| 1 | ResNet12 | 67.60 ± 0.44 | 66.83 ± 0.36 ⬇️ 📋 | 82.58 ± 0.30 | 82.13 ± 0.26 ⬇️ 📋 | Reproduce |
| 2 | Conv64F | - | 57.62 ± 0.36 ⬇️ 📋 | - | 74.14 ± 0.27 ⬇️ 📋 | Table.2 |
| 3 | ResNet12 | - | 64.81 ± 0.37 ⬇️ 📋 | - | 79.90 ± 0.27 ⬇️ 📋 | Table.2 |
| 3 | ResNet18 | - | 62.86 ± 0.37 ⬇️ 📋 | - | - | Table.2 |
| Embedding | 📖 tieredImageNet (5,1) | 💻 tieredImageNet (5,1) | 📖tieredImageNet (5,5) | 💻 tieredImageNet (5,5) | 📝 Comments | |
|---|---|---|---|---|---|---|
| 1 | ResNet12 | 71.61 ± 0.5 | 71.23 ± 0.42 ⬇️ 📋 | 85.28 ± 0.35 | 84.78 ± 0.29 ⬇️ 📋 | Reproduce |
| 2 | Conv64F | - | 61.62 ± 0.40 ⬇️ 📋 | - | 76.74 ± 0.33⬇️ 📋 | Table.2 |
| 3 | ResNet12 | - | 70.14 ± 0.43 ⬇️ 📋 | - | 82.70 ± 0.31 ⬇️ 📋 | Table.2 |
| 4 | ResNet18 | - | 71.53 ± 0.43 ⬇️ 📋 | - | - ⬇️ 📋 | Table.2 |