Image Classification on ELEVATER and ImageNet 21 multimodal datasets (test)
56.8AccuracyNegCLIP
Evaluation results
| Method | Links | |
|---|---|---|
| NegCLIPZero-shot=true, Aggregation=Average of 21 datasets2023.05 | 56.8 | |
| MosaiCLIP WiSE-FTZero-shot=true, Aggregation=Average of 21 datasets, Weight-Space Ensembled=true2023.05 | 56.8 | |
| CLIPZero-shot=true, Aggregation=Average of 21 datasets2023.05 | 56.4 | |
| MosaiCLIP NoCurricZero-shot=true, Aggregation=Average of 21 datasets, Curriculum Learning=false2023.05 | 55.8 | |
| MosaiCLIPZero-shot=true, Aggregation=Average of 21 datasets2023.05 | 55.7 |