Completed
Defining custom gradients
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Flux - The Elegant Machine Learning Library for Julia
Automatically move to the next video in the Classroom when playback concludes
- 1 Welcome!
- 2 Quick overview of Flux.jl
- 3 Quick glimpse on Flux.jl internals
- 4 Summer of Code with Flux.jl
- 5 Reading ONNX files with ONNX.jl
- 6 Images recognition models
- 7 Speech recognition model with CUDA
- 8 Demo of speech recognition
- 9 Reinforcement learning and AlphaGo.jl
- 10 Exporting Flux.jl to the browser with FluxJS.jl
- 11 Cheers for people from Summer of Code
- 12 Plans for the future: ML for us is a compiler problem
- 13 Our main compiler problem: automatic differentiation (AD)
- 14 AD normally needs expression trees
- 15 Can we do better? Our answer: Zygote.jl
- 16 Taking derivative and Julia IR (Internal Representation)
- 17 Benchmarking more complex examples
- 18 Speed, but at what cost?
- 19 Defining custom gradients
- 20 Convenient error messages
- 21 We have fully dynamics AD
- 22 In Julia, we can just hack compilers with different tricks when we need it
- 23 Demo of simple derivative
- 24 Question: what is that arrow?
- 25 Q&A: can you differential different functions that number to number (scalar to scalar)?
- 26 Comment: removing of the stack in some cases
- 27 Q&A: what are the Zygote.jl limitations right now?
- 28 Q&A: what is relation of Zygote.jl and Casset.jl?
- 29 Q&A: can we use Zygote.jl to differentiate function from one parameter to many parameters?
- 30 Annulment of Flux.jl on hackathon
