Image Retrieval on Animals with Attributes (AwA) unlabeled (Vut) (train)
74.2mAPReViSE
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ReViSETraining Data=labeled and unlabeled training data, Attribute Type=att, Backbone=GoogLeNet2017.03 | 74.2 | 57.6 | |
| ReViSEcTraining Data=labeled and unlabeled training data, Attribute Type=att, Backbone=GoogLeNet, Note=without unsupervised-data adaptation (beta=0)2017.03 | 66.8 | 51.8 | |
| DeViSE*Training Data=labeled and unlabeled training data, Attribute Type=att, Backbone=GoogLeNet, Note=with unsupervised-data adaptation2017.03 | 65.6 | 47.8 | |
| ReViSEbTraining Data=labeled training data only, Attribute Type=att, Backbone=GoogLeNet, Objective=supervised + unsupervised on labeled data2017.03 | 64.4 | 50.3 | |
| CMT*Training Data=labeled and unlabeled training data, Attribute Type=att, Backbone=GoogLeNet, Note=with unsupervised-data adaptation2017.03 | 63.5 | 49.6 | |
| DeViSETraining Data=labeled training data only, Attribute Type=att, Backbone=GoogLeNet2017.03 | 60.5 | 46.6 | |
| ReViSEaTraining Data=labeled training data only, Attribute Type=att, Backbone=GoogLeNet, Objective=supervised only (alpha=0)2017.03 | 60.2 | 46.4 | |
| CMTTraining Data=labeled training data only, Attribute Type=att, Backbone=GoogLeNet2017.03 | 58.8 | 48.3 |