Zero Shot Cross Lingual Transfer On Xtreme
Metrics
Avg
Question Answering
Sentence Retrieval
Sentence-pair Classification
Structured Prediction
Results
Performance results of various models on this benchmark
Comparison Table
| Model Name | Avg | Question Answering | Sentence Retrieval | Sentence-pair Classification | Structured Prediction |
|---|---|---|---|---|---|
| Model 1 | 56.1 | 68.6 | NA | 84.1 | 73.3 |
| Model 2 | 79.9 | 71.7 | 89.0 | 88.2 | 74.6 |
| ernie-m-enhanced-multilingual-representation | 80.9 | 72.3 | 91.9 | 87.9 | 75.6 |
| Model 4 | 53.1 | 52.5 | 18.0 | 75.3 | 66.9 |
| english-intermediate-task-training-improves | 73.5 | 67.2 | 76.5 | 83.9 | 69.4 |
| Model 6 | 81.6 | 72.5 | 93.7 | 88.4 | 76.2 |
| Model 7 | 81.7 | 71.9 | 90.8 | 88.3 | 80.6 |
| infoxlm-an-information-theoretic-framework | 80.7 | 72.9 | 89.3 | 88.8 | 75.4 |
| Model 9 | 76.5 | 59.7 | 77.5 | 86.3 | 90.8 |
| Model 10 | 82.0 | 73.4 | 93.3 | 89.0 | 76.7 |
| mt5-a-massively-multilingual-pre-trained-text | 40.9 | 73.6 | NA | 89.8 | NA |
| filter-an-enhanced-fusion-method-for-cross | 77.0 | 68.5 | 84.4 | 87.5 | 71.9 |
| Model 13 | 85.0 | 76.3 | 94.4 | 90.4 | 83.1 |
| Model 14 | 85.0 | 76.3 | 94.4 | 90.4 | 83.1 |
| xlm-e-cross-lingual-language-model-pre | 85.5 | 77.1 | 94.4 | 91.0 | 83.8 |
| Model 16 | 83.7 | 74.3 | 93.7 | 90.0 | 81.4 |
| Model 17 | 80.8 | 71.9 | 92.6 | 89.0 | 74.4 |
| Model 18 | 84.1 | 75.0 | 94.2 | 90.1 | 81.4 |
| xtreme-a-massively-multilingual-multi-task-1 | - | 53.8 | 47.7 | 73.7 | 66.3 |
| Model 20 | 82.2 | 75.2 | 92.4 | 89.3 | 75.5 |
| Model 21 | 82.0 | 75.2 | 92.4 | 89.2 | 74.6 |
| Model 22 | 39.3 | 65.5 | 34.5 | 44.2 | 0.0 |
| Model 23 | 82.0 | 73.4 | 93.3 | 89.0 | 76.7 |
| xlm-e-cross-lingual-language-model-pre | 84.5 | 76.3 | 93.7 | 90.3 | 81.7 |
| Model 25 | 81.6 | 72.5 | 93.7 | 88.4 | 76.2 |