|
|
|
# Self-Learning: Deep-Learning
|
|
|
|
|
|
|
|
## Goal
|
|
|
|
|
|
|
|
> The goal of this section is to dig into deep-learning in a practical way. In our opinion as an university of applied Sciences the combination of theoratical an practical content is very sustainable.
|
|
|
|
>
|
|
|
|
> After finishing this section you should have an idea how you can use datasets to train ML/AI Models and use them!
|
|
|
|
|
|
|
|
### Intro into Deep Learning
|
|
|
|
|
|
|
|
> In our practical session you´re going to develop your own prediction-model and hopefully hack your coffeemaschine.
|
|
|
|
>
|
|
|
|
> Therefore we we want you to accomplish a baisc Deep-Learning course:
|
|
|
|
>
|
|
|
|
> Kaggle.com is not only a great source for Datasets, it also provides good Learning Content.
|
|
|
|
>
|
|
|
|
> Daher bearbeitet bitte den folgenden Lernkurs: https://www.kaggle.com/learn/intro-to-deep-learning
|
|
|
|
>
|
|
|
|
> Am Ende des 6 lessons bekommt ihr ei n Certificate ausgestellt.
|
|
|
|
|
|
|
|
### Limitations
|
|
|
|
|
|
|
|
> As you can imagine a hugh Dataset, multiple layers, iterations are increasing the neccessary ressources for training the model dramatically. Thats one reason why the kaggle-course is implemented as an Notebook.
|
|
|
|
>
|
|
|
|
> A notebook provides you nothing else than running your Code not on your own hardware. But the code is running on external high performance server. For our project we will setup a different approach. Keeping the Datasets and model as a lighweight solutions allows you to run the model on your own lightweight hardware.
|
|
|
|
>
|
|
|
|
> The trend towards lightweight AI models is gaining momentum, driven by the need for deploying AI on resource-constrained devices such as smartphones and edge devices.
|
|
|
|
|
|
|
|
### |
|
|
\ No newline at end of file |
