Sentiment classification on amazon fr (test)
5.95Error Rate (%)XLM_ft (Monolingual Baseline, UDA)
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| XLM_ft (Monolingual Baseline, UDA)Base Model=XLM_ft, Setup=Monolingual, Train Protocol=UDA, Unlabeled=true2019.09 | 5.95 | — | — | — | — | |
| XLM_ft (UDA+Self)Base Model=XLM_ft, Train Protocol=UDA+Self, Unlabeled=true2019.09 | 6.67 | — | — | — | — | |
| XLM_ft (UDA)Base Model=XLM_ft, Train Protocol=UDA, Unlabeled=true2019.09 | 7.15 | — | — | — | — | |
| XLM_ft (Ft)Base Model=XLM_ft, Train Protocol=Ft, Unlabeled=true2019.09 | 7.67 | — | — | — | — | |
| XLM (Monolingual Baseline, Ft)Base Model=XLM, Setup=Monolingual, Train Protocol=Ft, Unlabeled=false2019.09 | 8.32 | — | — | — | — | |
| XLM (UDA+Self)Base Model=XLM, Train Protocol=UDA+Self, Unlabeled=true2019.09 | 9.03 | — | — | — | — | |
| XLM (UDA)Base Model=XLM, Train Protocol=UDA, Unlabeled=true2019.09 | 9.83 | — | — | — | — | |
| XLM (Ft)Base Model=XLM, Train Protocol=Ft, Unlabeled=false2019.09 | 11.35 | — | — | — | — | |
| CLDFAfine-tuned=true2021.05 | — | 83.37 | 82.56 | 83.31 | 83.08 | |
| CLHGfine-tuned=true2021.05 | — | 90.67 | 91.38 | 90.45 | 90.83 | |
| mBERTfine-tuned=true2021.05 | — | 84.55 | 85.85 | 83.65 | 84.68 | |
| MVECfine-tuned=true2021.05 | — | 89.08 | 88.28 | 88.5 | 88.62 | |
| XLMfine-tuned=true2021.05 | — | 88.1 | 86.95 | 86.2 | 87.08 | |
| XLM-Rfine-tuned=true2021.05 | — | 89.33 | 90.07 | 89.15 | 89.52 |