Generalized Few-Shot Image Classification on MiniImageNet (test)
65.33AccuracyFEAT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FEATShots=5-shot, Backbone=ConvNet2018.12 | 65.33 | — | — | — | |
| ProtoNetShots=5-shot, Backbone=ConvNet2018.12 | 64.94 | — | — | — | |
| FEATShots=1-shot, Backbone=ConvNet2018.12 | 49.72 | — | — | — | |
| FEATShots=5-shot, Backbone=ConvNet2018.12 | 44.94 | — | — | — | |
| FEATShots=1-shot, Backbone=ConvNet2018.12 | 43.94 | — | — | — | |
| ProtoNetShots=1-shot, Backbone=ConvNet2018.12 | 41.73 | — | — | — | |
| FEATShots=5-shot, Backbone=ConvNet2018.12 | 41.68 | — | — | — | |
| ProtoNetShots=5-shot, Backbone=ConvNet2018.12 | 41.06 | — | — | — | |
| FEATShots=1-shot, Backbone=ConvNet2018.12 | 40.5 | — | — | — | |
| ProtoNetShots=5-shot, Backbone=ConvNet2018.12 | 38.04 | — | — | — | |
| ProtoNetShots=1-shot, Backbone=ConvNet2018.12 | 35.69 | — | — | — | |
| Random Chance2018.12 | 20 | — | — | — | |
| ProtoNetShots=1-shot, Backbone=ConvNet2018.12 | 18.64 | — | — | — | |
| Random Chance2018.12 | 1.56 | — | — | — | |
| Random Chance2018.12 | 1.45 | — | — | — | |
| FEATtraining_protocol=1-shot 5-way2018.12 | — | 43.94 | 49.72 | 40.5 | |
| FEATtraining_setup=1-shot 5-way2018.12 | — | 43.94 | 49.72 | 40.5 | |
| ProtoNettraining_protocol=1-shot 5-way2018.12 | — | 41.73 | 48.64 | 35.69 | |
| ProtoNettraining_setup=1-shot 5-way2018.12 | — | 41.73 | 48.64 | 35.69 | |
| Randomtraining_protocol=1-shot 5-way2018.12 | — | 1.56 | 20 | 1.45 | |
| Randomtraining_setup=1-shot 5-way2018.12 | — | 1.56 | 20 | 1.45 |