A Japanese NLI Cross‑Encoder built on tohoku-nlp/bert-base-japanese-v3, trained on JSNLI, JNLI, and JSICK datasets, outputs entailment/neutral/contradiction scores with ~91.4 % accuracy on the JGLUE‑JNLI validation set.
Description in other languages
Statistics
Statistics as of Nov 16, 2026
Hugging Face downloads (last 30 days)
1,000
Hugging Face likes
5
Listed
May 13, 2024
Resources in similar research fields
- bert-finetuned-japanese-sentimentchristian-phu329kFine‑tuned Japanese BERT (cl‑tohoku/bert‑base‑japanese‑v2) on Amazon product reviews for sentiment classification, achieving ~81 % accuracy and 0.73 F1 after 6 epochs with a 2 × 10⁻⁵ learning rate.Shared: Sentiment & text classification・Pretrained encoders (BERT etc.)
- deberta-v2-base-japaneseizumi-lab3kDeBERTaV2 base trained on Japanese corpora (CC‑100, mC4, OSCAR2301, Wikipedia, Wikinews) with FP‑16 fine‑tuning for NLU tasks (JSTS, JNLI, JCommonsenseQA), released under CC BY‑SA 4.0 and funded by Japanese research grants.Shared: Pretrained encoders (BERT etc.)・QA & reasoning
BERT-Japan-vaccinationRamos-Ramos6Official fine-tuning code for "Emotion Analysis of Japanese Tweets and Comparison to Vaccinations in Japan"GitHubShared: Sentiment & text classification・Pretrained encoders (BERT etc.)- bert-base-japanese-v2-wrime-fine-tunepatrickramos1kA Japanese BERT BASE fine‑tuned on the WRIME dataset predicts 0‑4 intensity scores for eight emotions (joy, sadness, anticipation, surprise, anger, fear, disgust, trust) for writers and readers, with code available, trained in 3 hrs on a K80, achieving about 0.6 MSE for writers and 0.2 MSE for readers.Shared: Sentiment & text classification・Pretrained encoders (BERT etc.)
- bert-base-japanese-v3-jstsllm-book1kA Japanese BERT‑based model fine‑tuned on the JGLUE JSTS dataset for semantic similarity scoring—introduced in chapter 5 of “Large Language Model Introduction”—with Colab notebooks, transformers‑pipeline usage, and an Apache 2.0 license.Shared: Sentiment & text classification・Pretrained encoders (BERT etc.)
Japanese-BERT-Sentiment-Analyzerizuna3852Deploying sentiment analysis server with FastAPI and BERTGitHubShared: Sentiment & text classification・Pretrained encoders (BERT etc.)