From the course: Applied AI: Getting Started with Hugging Face Transformers
Unlock the full course today
Join today to access over 23,200 courses taught by industry experts.
Pretrained Transformer models
From the course: Applied AI: Getting Started with Hugging Face Transformers
Pretrained Transformer models
- [Instructor] Pre-trained transformer models are becoming popular these days, as they make building use cases quicker and easier. Pre-trained transformer models are general purpose models built based on the transformer architecture. They're also called foundational models and language models. They're built using a large corpora of data and shared by the community. They can be used to solve a wide variety of NLP tasks. What is a pre-trained transformer model? Pre-trained models have a specific architecture along with a trained set of parameters and hyper-parameters. When you download a pre-trained model, it comes with its model architecture definition, as well as its parameters like weights. Pre-trained transformers are usually created for tasks like masked language modeling, and next sentence prediction. Masked language modeling is done by randomly hiding a given percent of tokens from a sentence, and building a model to…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.