Any Machine learning technique requires you to have labelled data to train the model / build a classifier. But a lot of Business have no training or annotated data to efficiently build a Text Classification or Sentiment Classification Model. This new Zero-Shot Classification Pipeline from Hugging Face is a blessing for any such company. All you need to import is hugging face transformers library and use the pipeline `zero-shot-classification` to get your State-of-the-Art (SOTA) Text Classification Model up and running. It also can do Sentiment Classification into Positive and Negative. Extremely easy to use for any programmer.
Google Colab (by Joe Davison) - [ Ссылка ]
Zero-Shot Learning in Modern NLP - [ Ссылка ]
Hugging face 🤗 Transformers - [ Ссылка ]
Topics: zero-shot classification without annotated data, NLP without training data, NLP Text Classification using Deep Learning, Sentiment Classification using Deep Learning, Python Sentiment Analysis, Python NLP Classification
![](https://i.ytimg.com/vi/45rVF3t8OII/maxresdefault.jpg)