Cross Lingual Natural Language Inference On 4

Accuracy

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

		Paper Title
ByT5 XXL	83.7	ByT5: Towards a token-free future with pre-trained byte-to-byte models
Decoupled	71.3	Rethinking embedding coupling in pre-trained language models
Coupled	70.7	Rethinking embedding coupling in pre-trained language models
ByT5 Small	69.1	ByT5: Towards a token-free future with pre-trained byte-to-byte models
mGPT	40.6	mGPT: Few-Shot Learners Go Multilingual

0 of 5 row(s) selected.