This repository contains notebooks implementing the code samples found in the video course Deep Learning Crash Course (Manning Publications). Note that the video course features far more content than you will find in these notebooks, in particular further explanations and figures. Here we have only included the code samples themselves and immediately related surrounding comments.
In this course we train a model that can predict the crash risk of a driver based on three simple inputs. We train the model using Colab notebooks on Google's GPU based hardware and convert the final model to a format TensorFlow.js supports. This allows us to deploy the model togther with a simple application that runs serverless in the browser.
Try is out here (you will need an up to version of a modern browser to do this):
https://djcordhose.github.io/deep-learning-crash-course-notebooks/
These notebooks have been created using Python 3.6 and TensorFlow 1.x
-
Unit 3:
- Notebook 3.3: Getting to know our data
- Notebook 3.5: Introduction to Neural Networks with TensorFlow and Keras Layers
- Notebook 3.7: Training Neural Networks with TensorFlow and Keras Layers
- Exercise 3.7: Draw Decision Boundaries by hand
- Notebook 3.9: Regularization - How to get a decent architecture?
- Notebook 3.10: Our Final TensorFlow Model
-
Unit 4:
TensorFlow models from Unit 3 converted to TensorFlow 2
- Unit 3 converted to TensorFlow 2:
