| Barba is a multilingual [natural language inference](http://nlpprogress.com/english/natural_language_inference.html) model for [textual entailment](https://en.wikipedia.org/wiki/Textual_entailment) and [zero-shot text classification](https://joeddav.github.io/blog/2020/05/29/ZSL.html#Classification-as-Natural-Language-Inference), available as an end-to-end service through TensorFlow Serving. Based on [XLM-RoBERTa](https://arxiv.org/abs/1911.02116), it is trained on selected subsets of publicly available English ([GLUE](https://huggingface.co/datasets/glue)), Chinese ([CLUE](https://huggingface.co/datasets/clue)), Japanese ([JGLUE](https://huggingface.co/datasets/shunk031/JGLUE)), Korean ([KLUE](https://huggingface.co/datasets/klue)) datasets, as well as other private datasets. |
