Cross-modal Retrieval on MSCOCO (5K)
83.4Mean Recall (ja)ALIGN-L2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| ALIGN-L2Evaluation Protocol=Fine-tuned, Data=AT-en, Type=Dual Encoder2021.09 | 83.4 | — | |
| MURAL-LARGEEvaluation Protocol=Fine-tuned, Data=AT+MBT, Type=Dual Encoder2021.09 | 81.3 | 81.2 | |
| MURAL-BASEEvaluation Protocol=Fine-tuned, Data=AT+MBT, Type=Dual Encoder2021.09 | 74.9 | 75.4 | |
| ALIGN-BASEEvaluation Protocol=Fine-tuned, Data=AT, Type=Dual Encoder2021.09 | 74.1 | 76.1 | |
| ALIGN-BASEEvaluation Protocol=Fine-tuned, Data=TrTrain(AT-en), Type=Dual Encoder2021.09 | 72.5 | 74.8 | |
| MURAL-BASEEvaluation Protocol=Fine-tuned, Data=TrTrain(CC12m)+EOBT, Type=Dual Encoder2021.09 | 71.9 | 73.7 | |
| MURAL-LARGEEvaluation Protocol=Zero-shot, Data=AT+MBT, Type=Dual Encoder2021.09 | 64.6 | 67.7 | |
| MURAL-BASEEvaluation Protocol=Zero-shot, Data=AT+MBT, Type=Dual Encoder2021.09 | 54.4 | 59.5 | |
| ALIGN-BASEEvaluation Protocol=Zero-shot, Data=AT, Type=Dual Encoder2021.09 | 53.9 | 59.6 | |
| ALIGN-BASE-ENEvaluation Protocol=Zero-shot, Data=AT-en-translate-test, Type=Dual Encoder2021.09 | 51.9 | 60.6 | |
| MURAL-BASEEvaluation Protocol=Zero-shot, Data=TrTrain(CC12m)+EOBT, Type=Dual Encoder2021.09 | 49.7 | 58 | |
| ALIGN-BASEEvaluation Protocol=Zero-shot, Data=TrTrain(AT-en), Type=Dual Encoder2021.09 | 46 | 55.9 |