Topic classification with pretrained models in Python - Python Tutorial

- [Instructor] In this example, we're going to use a pre-trained model from the Hugging Face hub for topic classification. Topic classification is a technique used to assign predefined topics or categories such as weather, sports, finance and more to a given piece of text. Let's start by selecting a kernel. Then we minimize or reduce the verbosity of our logs. So now, we're ready to instantiate a pipeline for topic classification. So the Hugging Face hub does not have a dedicated pipeline named for topic classification. However, we can still use generic text classification pipeline for this purpose. But for this to occur, we would need to use a model specifically trained for topic classification. In this example, we're going to use the multilingual-IPTC-news-topic-classifier. To use this model, we begin by importing the pipeline function from the transformers package. We specify the name of our model and that we instantiate a new pipeline called topics. Within the pipeline function…