From the course: AI Workshop: Advanced Chatbot Development
Unlock this course with a free trial
Join today to access over 25,300 courses taught by industry experts.
Solution: Prepare the chatbot for deployment
From the course: AI Workshop: Advanced Chatbot Development
Solution: Prepare the chatbot for deployment
(bright upbeat music) - [Instructor] So how was the challenge? In this challenge, we had to create a chatbot of a Flan-T5 model after doing distillation and doing pruning. I know it's a lot, but it kind of, it's a nice twist to everything that we have seen in this chapter. So don't worry. Let's go with step by step, and as I always say, this is a solution, not the solution, there are many, many ways to do this, so if it doesn't look exactly as what you did, it's super, super fine. So let's connect to a GPU. There we are. Now we have our GPU, and as always we start with the pip installs. There it is, beautiful. So, first we are going to do the distillation. Now, as you are well aware by now, distillation is a process that takes a lot of RAM, RAM in the CPU and RAM in the GPU, which means that in order to make it all fit all the processes in just one collab, we're going to do a, let's call it quote unquote demo distillation. So a very tiny dataset, just a thousand training examples, and…
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.
Contents
-
-
-
-
-
(Locked)
Principles of model pruning5m 1s
-
(Locked)
Demo: Pruning the chatbot model8m 19s
-
(Locked)
Theory and practice of model distillation6m 58s
-
(Locked)
Demo: Applying model distillation to the chatbot8m 38s
-
(Locked)
Understanding and implementing quantization6m 34s
-
(Locked)
Demo: Quantizing the chatbot model5m 35s
-
(Locked)
Demo: Overview of the results10m 47s
-
(Locked)
Solution: Prepare the chatbot for deployment11m 12s
-
(Locked)
-
-
-
-