Skip to content

Latest commit

 

History

History
19 lines (18 loc) · 7.02 KB

wikiann_zero_shot_xlm_r_results.md

File metadata and controls

19 lines (18 loc) · 7.02 KB

Development Results:

Model Name ro gu pa lt az uk pl qu hu fi et tr kk zh my yo sw th ko ka ja ru bg es pt it fr fa ur mr hi bn el de en nl af te ta ml eu tl ms jv id vi he ar Avg.
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-1/best-model.pt 72.6 55.2 45.1 74.9 70.2 78.2 78.6 56.8 78.9 74.8 72.4 76.8 47.9 30.7 57 36.1 68.6 4.3 51.2 65.2 22.2 65.2 77.8 75.4 79 78.2 77.4 50.2 58.3 60.6 70.3 69.8 73.1 75.1 83.5 80.5 74.5 51.6 55.9 61 60.8 73.7 54.4 51.2 50.2 69.6 52.3 47.9 62.9
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-2/best-model.pt 72.1 52.7 44.6 74.9 70.4 70.2 78.9 56.1 78.1 74.9 72 76.1 46.8 28.4 57.7 31.4 68.6 4.2 49.2 68 21.3 64.4 78.8 76.4 78.9 78.4 77.7 47.4 59.1 62.2 67.6 68.2 74.8 75.5 83.7 80.4 76.1 48 57.4 61.1 59.4 71.6 56.5 52.1 52.3 66.2 51.4 48.6 62.4
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-3/best-model.pt 70.9 56.7 47.3 73.8 67.9 75.7 78.9 48 77 74.8 72.9 74.2 49.2 27.5 52.2 34 66.3 4.6 50.5 68.7 21 65.3 78.9 73.6 78.1 77.4 77.2 47.7 55.5 58.5 68.3 69.1 73.3 74.8 83.8 80.9 75 49.2 55.9 60.9 55.9 71 65.6 49 50.6 66.3 51.7 50.6 62.2
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-4/best-model.pt 73 54.3 44.7 74.9 69.1 76.9 79.2 57 78 75 73.1 76.4 44.2 28.7 55.4 35.1 69.4 5.1 50.3 65.5 21.2 65.5 78.2 77.6 77.9 77.9 77.2 47.1 53.3 60.1 68.1 69.3 73.2 75.2 83.7 80.6 75.9 49.7 56.1 57.4 58.4 70.1 65.7 52.1 50.4 67.1 52 44.5 62.3
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-5/best-model.pt 78.6 54.3 42.7 75 68.7 78.4 79.4 58.7 80.2 75.7 74.3 78.1 45 29.7 57.6 38.6 66.6 4.7 51.1 67.5 21 65.9 78.7 77.9 79.4 78.7 78.6 51.3 62.3 61.5 69.9 66.9 74.9 75.7 83.7 80.6 75.2 50.8 56.3 60.6 59.6 73.1 68.1 57.3 51.7 66.3 54.2 48.6 63.4
Language Avg. 73.4 54.6 44.9 74.7 69.3 75.9 79 55.3 78.4 75 72.9 76.3 46.6 29 56 35 67.9 4.6 50.5 67 21.3 65.3 78.5 76.2 78.7 78.1 77.6 48.7 57.7 60.6 68.8 68.7 73.9 75.3 83.7 80.6 75.3 49.9 56.3 60.2 58.8 71.9 62.1 52.3 51 67.1 52.3 48 62.2

Test Results:

Model Name ro gu pa lt az uk pl qu hu fi et tr kk zh my yo sw th ko ka ja ru bg es pt it fr fa ur mr hi bn el de en nl af te ta ml eu tl ms jv id vi he ar Avg.
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-1/best-model.pt 72.9 69.7 49.8 74.4 64.6 78.7 77.9 58.4 78.2 75.4 72.6 76.7 46.8 31.2 55.1 35.9 68.3 4.4 50.1 65.1 23.1 65 76.7 75.7 79.4 77.6 77.3 50.7 56 63 69.3 69.5 73.5 75.3 83.6 80.7 75.9 52.2 57.5 63.1 61.2 72.3 55.7 58.5 49.1 70 52.6 47.7 62.9
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-2/best-model.pt 72.7 67.1 50 74.7 65.7 70.8 78.8 60.2 77.3 75.9 72.5 76.3 45.1 29.4 51.3 37.8 68 4.3 48.1 68.5 21.8 64.5 78.2 77.3 79.1 77.9 78.2 47.9 56.4 61.5 66.5 70.2 74.8 75.3 83.4 80.4 76.4 46.7 58.2 62.2 59.7 71.8 58.6 56 51.3 66.9 51.5 48.2 62.4
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-3/best-model.pt 71.4 62.6 51.9 73.3 61.9 76.6 78.4 58.8 76.5 75.4 72.9 74.2 48 28.3 55 36.4 65.6 4.7 49.3 69.2 21.7 65.3 77.9 74.3 78.3 77.3 77.6 47.6 51.6 59.4 67.3 70.4 73.6 74.9 83.1 81 75.7 50 56 64.3 56.7 70.9 64.2 58.2 49.8 67.5 52.2 50.5 62.2
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-4/best-model.pt 73.3 61.8 53.6 74.3 62.9 77.5 78.7 64.4 77.2 75.9 72.9 76.6 42.3 29.2 50.7 39.8 68.4 5.4 49.1 66.4 21.4 65.4 77.1 78.1 78.3 77.2 77.4 47.1 51.4 60.6 67 69.7 73.8 75.5 83.5 80.7 76 47.9 56.1 59 58.6 73.1 65.8 57.3 49.2 67.9 52.1 44 62.3
wikiann-en-fine-tuned-xlm-roberta-base-bs32-wsFalse-e10-lr2e-05-layers-1-crfFalse-5/best-model.pt 78.6 66.2 47.7 74.6 65.1 79.1 78.4 62.1 79.5 76.8 74.2 78.2 43.6 30.8 49.6 38.3 64.7 4.7 50.3 68.1 21.3 65.8 77.6 78.3 80 78.6 78.9 51.9 60.2 61.9 68.8 68.2 74.9 76.1 83.5 81 75.2 49.9 56.3 62.4 59.5 72.7 67.4 61.6 50.8 67.1 54.5 48.7 63.4
Language Avg. 73.8 65.5 50.6 74.3 64 76.5 78.4 60.8 77.7 75.9 73 76.4 45.2 29.8 52.3 37.6 67 4.7 49.4 67.5 21.9 65.2 77.5 76.7 79 77.7 77.9 49 55.1 61.3 67.8 69.6 74.1 75.4 83.4 80.8 75.8 49.3 56.8 62.2 59.1 72.2 62.3 58.3 50 67.9 52.6 47.8 62.6