[Video event 82] Intro to FluxML and Machine Learning in Julia

[samvmdev]

Kyle Daruwalla: Intro to FluxML and Machine Learning in Julia

00:00 Data Umbrella introduction 
06:47 Kyle begins talk - Introduction
08:03 Introduction to Julia programming language
10:30 Why choose Julia for ML?
15:02 Getting familiar with Julia
16:19 Creating the first array (tensor)
17:30 Matrix and vector
21:32 Matrix and vector - audio cuts
21:40 Multiple dispatch begins - audio out
22:18 Multiple dispatch - audio returns
23:49 Multiple dispatch + array interfaces, “one hot array”
28:31 Understanding GPU programming
30:44 Introduction to FluxML
31:33 FluxML ecosystem
35:16 Linear regression - ground truth model & generating data
37:25 Linear regression - defining a model (fit to ground truth)
38:51 Linear regression - taking a gradient
40:35 Linear regression - updating parameters of original model
42:04 Layers in Flux.jl (for scaling up)
45:25 Q&A
48:25 A more complex example - MNIST dataset
51:08 MNIST dataset example - defining the model and loss function
52:37 MNIST dataset example - defining the accuracy metric
54:16 MNIST dataset example - writing the training loop
56:21 Next steps (documentation, contributing, JuliaCon)
59:25 Q&A

## Resources
https://docs.julialang.org/en/v1/stdlib/REPL/
https://cuda.juliagpu.org/stable/ 
https://fluxml.ai/Flux.jl/stable/ecosystem/ 
https://fluxml.ai/Flux.jl/stable/models/overview/ 
https://github.com/FluxML/model-zoo/blob/master/vision/mlp_mnist/mlp_mnist.jl 
https://fluxml.ai/Flux.jl/stable/ 
​​https://github.com/FluxML/Flux.jl/blob/master/CONTRIBUTING.md 
https://juliacon.org/2023/ 

[reshamas]

Thank you @greenhacks. The timestamps have been added to the video, the markdown page, and the checklist has been updated.