From the course: Advanced Guide to ChatGPT, Embeddings, and Other Large Language Models (LLMs)

Unlock this course with a free trial

Join today to access over 25,200 courses taught by industry experts.

Case study: Visual QA—Setting up parameters and data

Case study: Visual QA—Setting up parameters and data - ChatGPT Tutorial

From the course: Advanced Guide to ChatGPT, Embeddings, and Other Large Language Models (LLMs)

Start my 1-month free trial Buy for my team

Case study: Visual QA—Setting up parameters and data

“

- [Instructor] So for example, we have our own data collator here, which is going to collate data from our source, and our source of data, I will show you here. Our source of data is called the VQA system. Our VQA system has data that looks like this. We have an image which I have stored on my machine as a file path here because loading all of these images at once would not just crash my own laptop but crash most machines out there. But we have the file path for the image, we have the question asked for that image, what is this photo taken looking through. And then we have a set of answers from a crowd of people. So a bunch of people answered this question and we are going to take the images and question pairs that have a consensus of people agreeing on what the answer is because you can see here people wrote, net, net, net, netting, net, net, mesh, net, net, net. So arguably, there is a consensus here that net is the best answer, but not every image is going to have that consensus of…

Contents