Natural Language Visual Reasoning on NLVR2 std
85.5AccuracyX-VLM
Evaluation Results
| Method | Links | |
|---|---|---|
| X-VLMData=16M+, EPIC=true2022.11 | 85.5 | |
| METERData=16M+, EPIC=true2022.11 | 85.2 | |
| X-VLMData=4M+, EPIC=true2022.11 | 84.5 | |
| X-VLMData=16M+, EPIC=false2022.11 | 84.5 | |
| METERData=16M+, EPIC=false2022.11 | 84.3 | |
| X-VLMData=4M+, EPIC=false2022.11 | 84.1 | |
| ALBEFData=16M+, EPIC=true2022.11 | 84 | |
| METERData=4M+, EPIC=true2022.11 | 83.5 | |
| METERData=4M+, EPIC=false2022.11 | 82.5 | |
| ALBEFData=16M+, EPIC=false2022.11 | 82.5 | |
| ALBEFData=4M+, EPIC=true2022.11 | 82.2 | |
| ALBEFData=4M+, EPIC=false2022.11 | 80 | |
| ViLTData=4M, EPIC=true2022.11 | 77.1 | |
| ViLTData=4M, EPIC=false2022.11 | 75.2 |