TF Playground
This is a tutorial for better understanding the concept and syntax of Tensorflow. In this tutorial, we assume you already have the basic knowledge of Python and Machine Learning. You can fill the blank lines, run the script and see what would happen according to the code. This tutorial will begin with some style advice and some basic concepts, following by some exercises of different Machine Learnig models.
Please feel free to add some chapters if you think it might be good for our group.
Agenda
- Tensorflow Basics
- Deep Learning in TF
- Clustering
Sample Code
Here is a sample Tensorflow code, play around and grab as much information as you can here. We’ll explain it in detail in later chapters. When you think the code is ready, toggle the “run cell” button in the menu bar and see what will happen.
|
|
steps: 0, weight: 0.098388, bias: 0.440273
steps: 20, weight: 0.081874, bias: 0.310411
steps: 40, weight: 0.094683, bias: 0.303054
steps: 60, weight: 0.098440, bias: 0.300896
steps: 80, weight: 0.099542, bias: 0.300263
steps: 100, weight: 0.099866, bias: 0.300077
steps: 120, weight: 0.099961, bias: 0.300023
steps: 140, weight: 0.099988, bias: 0.300007
steps: 160, weight: 0.099997, bias: 0.300002
steps: 180, weight: 0.099999, bias: 0.300001
steps: 200, weight: 0.100000, bias: 0.300000
In the above sample, we defined a set of random inputs with y = 0.1 * x + 0.3
and let our linear model y = W * x + b
to learn the inputs in 200 training epochs. Results are printed per 20 time steps. W
and b
should be quite close to 0.1 and 0.3 respectively.