Completed
Creating image patches to tokenized
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Building Generalist Robotics Policies from Scratch
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro: ChatGPT, Language Models and the Goals of Generalist Robotics Policies
- 2 Reading and exploring the data
- 3 Creating a Dataset
- 4 Creating a Dataset
- 5 Creating the transformer encoder
- 6 Creating image patches to tokenized
- 7 Putting together the VIT
- 8 Training the VIT
- 9 Making the GRP, starting with adding text inputs
- 10 Modifying the data for training
- 11 Converting continuous actions to discrete bins
- 12 Converting continuous actions to discrete bins
- 13 Standardizing the state inputs
- 14 Changing to use continuous actions
- 15 Standizing the action space
- 16 Adding goal images to the transformer
- 17 Adding blocked masked attention to use either goal
- 18 Scaling training
- 19 Training results across A100s
- 20 Evaluation using the SimpleEnv robotics simulator
