# Machine Learning

**URL:** https://discourse.julialang.org/c/domain/ml/24.md?page=52

[Latest](https://discourse.julialang.org/latest.md) · [Categories](https://discourse.julialang.org/categories.md) · [Tags](https://discourse.julialang.org/tags.md)

**Page:** 53

---

## [Basics steps to build an ANN?](https://discourse.julialang.org/t/basics-steps-to-build-an-ann/21366)

<div class="topic-metadata">

**Author:** [@BLI](https://discourse.julialang.org/u/BLI)\
**Replies:** 24\
**Last updated:** [March 18, 2019, 10:04am UTC](https://discourse.julialang.org/t/basics-steps-to-build-an-ann/21366 "2019-03-18T10:04:38Z")

</div>

Is there a simple example on how to build a first ANN? I have a basic understanding of ANNs. I’m not looking for impressive examples. So, I’m not interested in hand writing recognition, image recognition, etc. Instead: …

---

## [Function in Flux to estimate learning rate](https://discourse.julialang.org/t/function-in-flux-to-estimate-learning-rate/21734)

<div class="topic-metadata">

**Author:** [@samq](https://discourse.julialang.org/u/samq)\
**Replies:** 7\
**Last updated:** [March 12, 2019, 9:36pm UTC](https://discourse.julialang.org/t/function-in-flux-to-estimate-learning-rate/21734 "2019-03-12T21:36:39Z")

</div>

Is there a function in Flux to estimate the best learning rate for a good gradient descent before training a neural network?

---

## [Flux results not similar to Tensorflow](https://discourse.julialang.org/t/flux-results-not-similar-to-tensorflow/21677)

<div class="topic-metadata">

**Author:** [@GJBoth](https://discourse.julialang.org/u/GJBoth)\
**Replies:** 3\
**Last updated:** [March 11, 2019, 9:26am UTC](https://discourse.julialang.org/t/flux-results-not-similar-to-tensorflow/21677 "2019-03-11T09:26:15Z")

</div>

Hi everyone, I currently do a lot of ML in Python using Tensorflow, which is working fine but Julia seems to be a bit more then fine, so I’m experimenting with Flux as an alternative. My first step is just to train a si…

---

## [Get in touch with gsoc mentors](https://discourse.julialang.org/t/get-in-touch-with-gsoc-mentors/21428)

<div class="topic-metadata">

**Author:** [@yash](https://discourse.julialang.org/u/yash)\
**Replies:** 7\
**Last updated:** [March 8, 2019, 2:20pm UTC](https://discourse.julialang.org/t/get-in-touch-with-gsoc-mentors/21428 "2019-03-08T14:20:42Z")

</div>

I am trying for gsoc this year in Julia projects in machine learning side. How can I get in touch with mentors?? And please guide me for my future steps.

---

## [Julia TextAnalysis](https://discourse.julialang.org/t/julia-textanalysis/21619)

<div class="topic-metadata">

**Author:** [@yash](https://discourse.julialang.org/u/yash)\
**Replies:** 2\
**Last updated:** [March 8, 2019, 8:42am UTC](https://discourse.julialang.org/t/julia-textanalysis/21619 "2019-03-08T08:42:01Z")

</div>

What does “rwi” parameter represents in get\_sentiment function in sentiment.jl file in TextAnalysis package?? Text Analysis : GitHub - JuliaText/TextAnalysis.jl: Julia package for text analysis

---

## [Automatic Differentiation over Vectors](https://discourse.julialang.org/t/automatic-differentiation-over-vectors/21603)

<div class="topic-metadata">

**Author:** [@benkj](https://discourse.julialang.org/u/benkj)\
**Replies:** 1\
**Last updated:** [March 7, 2019, 11:30pm UTC](https://discourse.julialang.org/t/automatic-differentiation-over-vectors/21603 "2019-03-07T23:30:01Z")

</div>

Consider the following expression c=exp(A\*t)\*b where A is a matrix, b and c are vectors and t is a scalar. I’m trying to teach ForwardDiff to calculate the derivative of c with respect to t, which is nothing but A\*c. Thi…

---

## [Deep forest NLP](https://discourse.julialang.org/t/deep-forest-nlp/21526)

<div class="topic-metadata">

**Author:** [@yash](https://discourse.julialang.org/u/yash)\
**Replies:** 6\
**Last updated:** [March 6, 2019, 11:43am UTC](https://discourse.julialang.org/t/deep-forest-nlp/21526 "2019-03-06T11:43:39Z")

</div>

Is there any Deep forest implementation in julia.???

---

## [Bug? Using Flux & getting Float32 response on 64 bit Ubuntu OS](https://discourse.julialang.org/t/bug-using-flux-getting-float32-response-on-64-bit-ubuntu-os/21440)

<div class="topic-metadata">

**Author:** [@samq](https://discourse.julialang.org/u/samq)\
**Replies:** 4\
**Last updated:** [March 4, 2019, 6:53pm UTC](https://discourse.julialang.org/t/bug-using-flux-getting-float32-response-on-64-bit-ubuntu-os/21440 "2019-03-04T18:53:25Z")

</div>

Julia version 1.1.0 Ubuntu 16.04 64 bit Flux version (?) please tell me how to see the version of Flux I get Float32 responses instead of Float64 when doing the following: using Flux model = Flux.Dense(2, 1, σ) mod…

---

## [Example of simple 1 layer Perception in Flux](https://discourse.julialang.org/t/example-of-simple-1-layer-perception-in-flux/21413)

<div class="topic-metadata">

**Author:** [@samq](https://discourse.julialang.org/u/samq)\
**Replies:** 8\
**Last updated:** [March 4, 2019, 6:38pm UTC](https://discourse.julialang.org/t/example-of-simple-1-layer-perception-in-flux/21413 "2019-03-04T18:38:33Z")

</div>

I saw the Flux pages, but the examples are not clear to me. Since Flux is simple and fast, would someone please post an example of how to make a simple 1 layer Perceptron using Flux? Such an example would be a springbo…

---

## [Why Julia machine learning is so unfriendly? Very "unsmooth" experience from foolish guy](https://discourse.julialang.org/t/why-julia-machine-learning-is-so-unfriendly-very-unsmooth-experience-from-foolish-guy/21174)

<div class="topic-metadata">

**Author:** [@kirtsar](https://discourse.julialang.org/u/kirtsar)\
**Replies:** 15\
**Last updated:** [March 3, 2019, 4:19pm UTC](https://discourse.julialang.org/t/why-julia-machine-learning-is-so-unfriendly-very-unsmooth-experience-from-foolish-guy/21174 "2019-03-03T16:19:01Z")

</div>

Ok, so let’s imagine that I’m a newcomer in a field of machine learning, and the very first task i want to do in Julia is something like knn-classifier. Nice! So let’s try to do it step by step, using all those cool libr…

---

## [\[DiffEqFlux\] How to do composition of layers?](https://discourse.julialang.org/t/diffeqflux-how-to-do-composition-of-layers/21291)

<div class="topic-metadata">

**Author:** [@pochoi](https://discourse.julialang.org/u/pochoi)\
**Replies:** 4\
**Last updated:** [March 2, 2019, 10:16pm UTC](https://discourse.julialang.org/t/diffeqflux-how-to-do-composition-of-layers/21291 "2019-03-02T22:16:03Z")

</div>

How to do composition of layers in DiffEqFlux? For example, I want do composite a layer of neural ODE with a layer of NN: dudt = Chain(Dense(2,40,tanh), Dense(40,2)) tspan = (0.0f0,1.0f0) t = range(tspan\[1\],tspan\[2\],le…

---

## [DiffEqFlux: how do declare a system of ODE which is backpropagatable?](https://discourse.julialang.org/t/diffeqflux-how-do-declare-a-system-of-ode-which-is-backpropagatable/21254)

<div class="topic-metadata">

**Author:** [@pochoi](https://discourse.julialang.org/u/pochoi)\
**Replies:** 2\
**Last updated:** [February 28, 2019, 5:58am UTC](https://discourse.julialang.org/t/diffeqflux-how-do-declare-a-system-of-ode-which-is-backpropagatable/21254 "2019-02-28T05:58:38Z")

</div>

Here is a simple example. Flux’s parameter seems to be not so friendly with matrix multiplication. When doing p .\* A where p is a parameter and A is a matrix, things go bad… using DifferentialEquations using Flux, Dif…

---

## [Group KFold?](https://discourse.julialang.org/t/group-kfold/21165)

<div class="topic-metadata">

**Author:** [@charlesll](https://discourse.julialang.org/u/charlesll)\
**Replies:** 4\
**Last updated:** [February 27, 2019, 10:30pm UTC](https://discourse.julialang.org/t/group-kfold/21165 "2019-02-27T22:30:46Z")

</div>

Hi, I am trying to implement a Group KFold function for my problem, where I need to slit the dataset as a function of some label. There is no function like that in MLDataUtils or MLbase to my knowledge, so I am trying …

---

## [Flux.jl and the state of multi-processing](https://discourse.julialang.org/t/flux-jl-and-the-state-of-multi-processing/21170)

<div class="topic-metadata">

**Author:** [@Alexander-Barth](https://discourse.julialang.org/u/Alexander-Barth)\
**Replies:** 2\
**Last updated:** [February 27, 2019, 7:51am UTC](https://discourse.julialang.org/t/flux-jl-and-the-state-of-multi-processing/21170 "2019-02-27T07:51:56Z")

</div>

I try to implement a convolutional auto encoder in Flux. The model is composed essentially of several convolutional layers, max pool, interpolation to the nearest neighborhood (using repeat(...; inner = (2,2,1,1)). The …

---

## [Flux not able to resolve strong observation, based on standard example that works](https://discourse.julialang.org/t/flux-not-able-to-resolve-strong-observation-based-on-standard-example-that-works/21177)

<div class="topic-metadata">

**Author:** [@ianshmean](https://discourse.julialang.org/u/ianshmean)\
**Replies:** 2\
**Last updated:** [February 25, 2019, 6:09pm UTC](https://discourse.julialang.org/t/flux-not-able-to-resolve-strong-observation-based-on-standard-example-that-works/21177 "2019-02-25T18:09:14Z")

</div>

I don’t understand why the housing example found here, which functions as expected, is not able to train on alternative data where I’m intentionally including the target in the observations to make it really easy. using…

---

## [Flux custom loss function](https://discourse.julialang.org/t/flux-custom-loss-function/21127)

<div class="topic-metadata">

**Author:** [@jstrube](https://discourse.julialang.org/u/jstrube)\
**Replies:** 3\
**Last updated:** [February 24, 2019, 9:45pm UTC](https://discourse.julialang.org/t/flux-custom-loss-function/21127 "2019-02-24T21:45:11Z")

</div>

I’m trying to run a straightforward regression problem through Flux. I’ve modified the loss function from the example, from a simple mse to something like a chi-square. The problem has to do with the calibration of a cal…

---

## [ADAM optimizer error in Flux](https://discourse.julialang.org/t/adam-optimizer-error-in-flux/20947)

<div class="topic-metadata">

**Author:** [@microlifecc](https://discourse.julialang.org/u/microlifecc)\
**Replies:** 1\
**Last updated:** [February 19, 2019, 6:01pm UTC](https://discourse.julialang.org/t/adam-optimizer-error-in-flux/20947 "2019-02-19T18:01:23Z")

</div>

Hi! I am defining a neural network as: n\_hidden = 20 model = Chain(Dense(n\_inputs, n\_hidden, relu), Dense(n\_hidden, n\_outputs, identity), softmax) L(x,y) = Flux.crossentropy(model(x), y) ps = Flux.params(…

---

## [Counterfactual Regret Minimization](https://discourse.julialang.org/t/counterfactual-regret-minimization/19543)

<div class="topic-metadata">

**Author:** [@ajkeith](https://discourse.julialang.org/u/ajkeith)\
**Replies:** 1\
**Last updated:** [February 15, 2019, 8:52pm UTC](https://discourse.julialang.org/t/counterfactual-regret-minimization/19543 "2019-02-15T20:52:11Z")

</div>

Is anyone aware of a Julia implementation of Counterfactual Regret Minimization (CFR)? I found a python implementation but I haven’t been able to find anything in Julia. I’m thinking about writing a Julia implementation…

---

## [Problem defining a Straight Through Estimator gradient](https://discourse.julialang.org/t/problem-defining-a-straight-through-estimator-gradient/20698)

<div class="topic-metadata">

**Author:** [@velix](https://discourse.julialang.org/u/velix)\
**Replies:** 0\
**Last updated:** [February 12, 2019, 11:31am UTC](https://discourse.julialang.org/t/problem-defining-a-straight-through-estimator-gradient/20698 "2019-02-12T11:31:00Z")

</div>

Hi all, I am trying to define a custom STE gradient in Flux. The activation function is the sign function, and its gradient should just be the incoming gradient as is, iff its absolute value is \<= 1, or cancelled other…

---

## [Understanding GRU forward pass as implemented in Knet](https://discourse.julialang.org/t/understanding-gru-forward-pass-as-implemented-in-knet/20639)

<div class="topic-metadata">

**Author:** [@Christian\_Dengler](https://discourse.julialang.org/u/Christian_Dengler)\
**Replies:** 1\
**Last updated:** [February 12, 2019, 5:00am UTC](https://discourse.julialang.org/t/understanding-gru-forward-pass-as-implemented-in-knet/20639 "2019-02-12T05:00:06Z")

</div>

Hello all, I tried to reimplement the simple forward pass of a GRU layer in order to be able to evaluate the RNN in other languages after training in Julia. So far with no luck, and Im starting to be out of ideas. Is th…

---

## [Automatic Differentiation](https://discourse.julialang.org/t/automatic-differentiation/17472)

<div class="topic-metadata">

**Author:** [@kkmann](https://discourse.julialang.org/u/kkmann)\
**Replies:** 11\
**Last updated:** [February 11, 2019, 2:19pm UTC](https://discourse.julialang.org/t/automatic-differentiation/17472 "2019-02-11T14:19:18Z")

</div>

Hey, I am still having a bit of a hard time figuring out what flux.jl actually does. One important aspect seems to be making gradients available automatically via tracking. How does that differ from the approach in Forw…

---

## [Pipeline : from raw data to fitted model?](https://discourse.julialang.org/t/pipeline-from-raw-data-to-fitted-model/20477)

<div class="topic-metadata">

**Author:** [@kirtsar](https://discourse.julialang.org/u/kirtsar)\
**Replies:** 6\
**Last updated:** [February 6, 2019, 10:22pm UTC](https://discourse.julialang.org/t/pipeline-from-raw-data-to-fitted-model/20477 "2019-02-06T22:22:31Z")

</div>

Hello everyone! I’m a novice in machine learning domain. I’m interested in basic ML pipeline from raw dataframe with missing values and so on to the trained model. Is there any unified process of pipelining such method…

---

## [Any Black-Box Packages for Bayesian Hyperparameter Optimization?](https://discourse.julialang.org/t/any-black-box-packages-for-bayesian-hyperparameter-optimization/20428)

<div class="topic-metadata">

**Author:** [@ahmetumutdurmus](https://discourse.julialang.org/u/ahmetumutdurmus)\
**Replies:** 2\
**Last updated:** [February 5, 2019, 9:42am UTC](https://discourse.julialang.org/t/any-black-box-packages-for-bayesian-hyperparameter-optimization/20428 "2019-02-05T09:42:55Z")

</div>

Configuring hyperparameters in a machine learning problem can be a dauntingly boring and suboptimal task as each evaluation may be fairly long to conduct and thus performing an exhaustive grid search may be infeasible. B…

---

## [Web site with deep learning examples](https://discourse.julialang.org/t/web-site-with-deep-learning-examples/20230)

<div class="topic-metadata">

**Author:** [@johnh](https://discourse.julialang.org/u/johnh)\
**Replies:** 3\
**Last updated:** [January 30, 2019, 10:15am UTC](https://discourse.julialang.org/t/web-site-with-deep-learning-examples/20230 "2019-01-30T10:15:00Z")

</div>

Sorry to ask a simple question. There is a website with live examples of Julia machine learning. I have searched but cannot find it. If I recall it is a .io site IF this rings a bell with anyone please let me have the a…

---

## [Parallel data loading to GPU arrays](https://discourse.julialang.org/t/parallel-data-loading-to-gpu-arrays/11321)

<div class="topic-metadata">

**Author:** [@kolia](https://discourse.julialang.org/u/kolia)\
**Replies:** 3\
**Last updated:** [January 30, 2019, 3:07am UTC](https://discourse.julialang.org/t/parallel-data-loading-to-gpu-arrays/11321 "2019-01-30T03:07:54Z")

</div>

Anyone have pointers to loading data into CuArrays with background tasks while the main thread or process is busy training a Flux model? Flux seems to have the same if not more flexibility than pytorch, so they fall in …

---

## [ANN: Knet v1.2.0: iterators, iterators, iterators](https://discourse.julialang.org/t/ann-knet-v1-2-0-iterators-iterators-iterators/19905)

<div class="topic-metadata">

**Author:** [@denizyuret](https://discourse.julialang.org/u/denizyuret)\
**Replies:** 8\
**Last updated:** [January 27, 2019, 10:24pm UTC](https://discourse.julialang.org/t/ann-knet-v1-2-0-iterators-iterators-iterators/19905 "2019-01-27T22:24:59Z")

</div>

The new Knet release is all about iterators: iterators for minibatching, iterators for training, iterators for monitoring, convergence etc. Why am I so excited about iterators all of a sudden? Allow me to explain: Knet …

---

## [DiffEqFlux.jl – A Julia Library for Neural Differential Equations](https://discourse.julialang.org/t/diffeqflux-jl-a-julia-library-for-neural-differential-equations/20000)

<div class="topic-metadata">

**Author:** [@MikeInnes](https://discourse.julialang.org/u/MikeInnes)\
**Replies:** 1\
**Last updated:** [January 24, 2019, 6:40am UTC](https://discourse.julialang.org/t/diffeqflux-jl-a-julia-library-for-neural-differential-equations/20000 "2019-01-24T06:40:42Z")

</div>

We just released a new blog post showing how you can replicate the results of the recent Neural ODEs paper – and then some – using Flux and DifferentialEquations.jl. We’ve been combining ML and DiffEqs for a quite some t…

---

## [What does \`Flux.@treelike\` do?](https://discourse.julialang.org/t/what-does-flux-treelike-do/19975)

<div class="topic-metadata">

**Author:** [@xiaodai](https://discourse.julialang.org/u/xiaodai)\
**Replies:** 2\
**Last updated:** [January 24, 2019, 3:24am UTC](https://discourse.julialang.org/t/what-does-flux-treelike-do/19975 "2019-01-24T03:24:12Z")

</div>

What does Flux.@treelike do? In Flux’s source code, I see a lot of Flux’s layers are defined and then @treelike is called. .E.g. @treelike Dense.

---

## [Deploy Julia model in flask](https://discourse.julialang.org/t/deploy-julia-model-in-flask/19961)

<div class="topic-metadata">

**Author:** [@VINITPANDEY](https://discourse.julialang.org/u/VINITPANDEY)\
**Replies:** 0\
**Last updated:** [January 23, 2019, 5:36am UTC](https://discourse.julialang.org/t/deploy-julia-model-in-flask/19961 "2019-01-23T05:36:05Z")

</div>

How to deploy machine learning model in Julia

---

## [Anybody working on Automatic Differentiation "the Conal Elliot way"?](https://discourse.julialang.org/t/anybody-working-on-automatic-differentiation-the-conal-elliot-way/19887)

<div class="topic-metadata">

**Author:** [@jbrea](https://discourse.julialang.org/u/jbrea)\
**Replies:** 5\
**Last updated:** [January 23, 2019, 12:11am UTC](https://discourse.julialang.org/t/anybody-working-on-automatic-differentiation-the-conal-elliot-way/19887 "2019-01-23T00:11:28Z")

</div>

Just learned about The simple essence of automatic differentiation. Did somebody already try to implement these ideas in Julia? Any thoughts about pros and cons of this approach relative to e.g. Zygote.jl?

[Previous page](https://discourse.julialang.org/c/domain/ml/24.md?page=51)

[Next page](https://discourse.julialang.org/c/domain/ml/24.md?page=53)