Predicting object attributes on PIGPEN-NLU unseen (test)
80.2AccuracyPIGLET
Evaluation results
| Method | Links | |
|---|---|---|
| PIGLET2021.06 | 80.2 | |
| T5-11BApproach category=text-to-text2021.06 | 59.1 | |
| T5-3BApproach category=text-to-text2021.06 | 58.7 | |
| T5-LargeApproach category=text-to-text2021.06 | 49.1 | |
| T5-BaseApproach category=text-to-text2021.06 | 48.8 | |
| Alberti et al.2019Approach category=BERT-style, Dynamics pre-training=true2021.06 | 48.1 | |
| T5-SmallApproach category=text-to-text2021.06 | 38 | |
| G&D2019Approach category=BERT-style, Dynamics pre-training=true2021.06 | 26.9 | |
| No Change2021.06 | 24 | |
| GPT3-175Bfew-shot=true2021.06 | 21.4 | |
| G&D2019Approach category=BERT-style, Dynamics pre-training=false2021.06 | 7.3 | |
| Alberti et al.2019Approach category=BERT-style, Dynamics pre-training=false2021.06 | 3.7 |