Advanced AI: Transformers for Computer Vision

55mAdvanced2023-03-03

Authors

Jonathan Fernandes

Consultant focusing on data science, AI, and big data

Course details

Transformers are quickly becoming the go-to architecture for many computer vision tasks. If you work in the field, it’s a must-have skill to keep on hand in your AI toolkit. In this course, AI consultant Jonathan Fernandes takes you on a deep dive into the world of transfer learning and transformer model architecture.

Explore the basics of computer vision, image datasets, preprocessing, and image fine-tuning, with hands-on examples and easy-to-follow demonstrations using Google Colab and the Hugging Face library. Discover tips and practical strategies for model training and testing as you go, building out your skill set with the popular inference modeling tools Gradio and Hugging Face Spaces. By the end of this course, you’ll be prepared to design and train larger, more advanced, more sophisticated language models.

Skills covered

PyTorchAdvancedGenerative AIPythonArtificial Intelligence (AI)Open Source

Concepts

0. Introduction

01 - Transformers for computer vision
02 - What you should know

1. Transformer Architecture

03 - History of transformers
04 - Comparing Vision Transformers to BERT

2. Datasets and Preprocessing

05 - Getting set up
06 - Getting the data
07 - Using datasets
08 - Using a pretrained model without fine-tuning
09 - Defining a model
10 - Preprocessing images
11 - A transformed image
12 - Getting images in the correct format

3. Model Training

13 - Training arguments
14 - Model training
15 - Inference in notebook
16 - Inference on phone using Gradio
17 - Gradio and Hugging Face Spaces

Conclusion

18 - Learn more about transformers and large language models