# First steps into Machine Learning?

**URL:** <https://discourse.julialang.org/t/first-steps-into-machine-learning/79153>\
**Category:** General Usage\
**Created:** [April 7, 2022, 10:05am UTC](https://discourse.julialang.org/t/first-steps-into-machine-learning/79153 "2022-04-07T10:05:05Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![marianoarnaiz](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/marianoarnaiz/32/19377_2.png) [@marianoarnaiz](https://discourse.julialang.org/u/marianoarnaiz)\
**Post date:** [April 7, 2022, 10:05am UTC](https://discourse.julialang.org/t/first-steps-into-machine-learning/79153/1 "2022-04-07T10:05:05Z")

</div>

Hi everyone!.

So … I am a ML newbie (I actually have just finished reading Jeremy Watt’s book and was able to understand some of it). Now I would like to do some computational testing, for example I have a task I need to get done and a dataset that might server to train a simple ML algorithm to do it for me. But, where should I begin with Julia?

Any tips and recommendations are welcome.

---

<div class="post-metadata">

**Author:** ![jmair](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/jmair/32/35117_2.png) [@jmair](https://discourse.julialang.org/u/jmair)\
**Post date:** [April 7, 2022, 10:20am UTC](https://discourse.julialang.org/t/first-steps-into-machine-learning/79153/2 "2022-04-07T10:20:19Z")

</div>

Check out some of the machine learning packages like Flux.jl and MLDatasets.jl. The first lets you easily write neural networks (and also contains ways of building loss functions). The second provides common datasets such as MNIST, which just gives you the data in a native tensor.

From just these you can train a simple neural network to classify handwritten digits (MNIST).

---

<div class="post-metadata">

**Author:** ![gdalle](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/gdalle/32/27854_2.png) [@gdalle](https://discourse.julialang.org/u/gdalle)\
**Post date:** [April 7, 2022, 11:08am UTC](https://discourse.julialang.org/t/first-steps-into-machine-learning/79153/3 "2022-04-07T11:08:48Z")

</div>

If you want to experiment with general ML algorithms, then [MLJ.jl](https://github.com/alan-turing-institute/MLJ.jl) is your best option.  
If you’re specifically interested in neural networks, then you should indeed check out [Flux.jl](https://github.com/FluxML/Flux.jl)

---

<div class="post-metadata">

**Author:** ![dmolina](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/dmolina/32/5246_2.png) [@dmolina](https://discourse.julialang.org/u/dmolina)\
**Post date:** [April 7, 2022, 3:53pm UTC](https://discourse.julialang.org/t/first-steps-into-machine-learning/79153/4 "2022-04-07T15:53:59Z")

</div>

I agree with @gdalle, only one addition. For using Flux.jl, it is useful to use also the higher-level functions [FastAI.jl](https://github.com/FluxML/FastAI.jl)

---

<div class="post-metadata">

**Author:** ![marianoarnaiz](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/marianoarnaiz/32/19377_2.png) [@marianoarnaiz](https://discourse.julialang.org/u/marianoarnaiz)\
**Post date:** [April 7, 2022, 8:25pm UTC](https://discourse.julialang.org/t/first-steps-into-machine-learning/79153/5 "2022-04-07T20:25:38Z")

</div>

> [@gdalle](#):
>
> Flux.jl

Thanks every one. I think I will begin with " An Introduction to Statistical Learning" and the examples on MLJ.

Thanks a lot!