From the course: Developing Modern Applications with AWS AI and Generative AI Services
Join today to access over 24,300 courses taught by industry experts.
Architecture behind generative AI - Amazon Web Services (AWS) Tutorial
From the course: Developing Modern Applications with AWS AI and Generative AI Services
Architecture behind generative AI
- [Instructor] We learned in the previous lesson that LLMs, a short form for large language models, are a type of foundation models. They are called large because they are trained with a large corpus of text data in the order of petabytes and billions of parameters. The second word, language, implies that these models are designed specifically for language processing tasks. LLMs are based on transformer models, which are, again, a type of neural network. Let's explore the architecture at a high level by breaking down into six different steps. The first step is Tokenization, where a sentence is broken into multiple tokens. This step takes every word in a sentence, including punctuation signs, prefixes and suffixes, and converts them into a token. For example, the sentence I am learning is converted into four different tokens, including the period at the end. The next step is embedding where each token is converted into numbers that the transformer model can understand. These numbers…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.