ML Introductory Practice Using Tensorflow
- tf.js is 3-5x slower than GPU tf
- define loss() or error()
- of predicted output to actual output
- checks parameter space, to min
- create linear regression
- apply stochastic descent
- loss() and confusion matrices
Obs | Instruction | Tasks |
---|---|---|
1. | Clone Ryan Marchildon's repo | $ git clone git@github.com:RyanMarchildon/tfjs-torontoai-lecture.git |
2-a. | Webserver - JS | $ npm install local-web-server -g |
2-b. | Webserver - Py | $ python3 -m http.server |
-
Fundamental
-
tf.js /index.html
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs@1.0.0/dist/tf.min.js"></script>
-
-
Tensors
- multi-dimensional array vector
-
CNN
- tfjs-vis
-
Memory management
-
MNIST data with tf.js
-
loss( ) and confusion matrices
-
local host this model due to cross origin reference sharing
-
need to add script call below in html file
<!-- Import TensorFlow.js --> <script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs@1.0.0/dist/tf.min.js"></script> <!-- Import tfjs-vis --> <script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs-vis@1.0.2/dist/tfjs-vis.umd.min.js"></script> <!-- Import the data file --> <script src="data.js" type="module"></script> <!-- Import the main script file --> <script src="main.js" type="module"></script>
-
NIP's Visual Reasoning by Progressive Module Networks - Seung Wook Kim
Rangle.io's take on building AI applications with tensorflow.js