Skip to content

peremartra/TensorFlow_Advanced

Repository files navigation

TensorFlow Advanced Techniques

It is a series of notebooks about some “beyond the basics” or mid-advanced techniques with TensorFlow and Keras.

It could be said that these notebooks pick up where the TensorFlow Developer Certified exam leaves off.

It is not a course or tutoria, just a colletion of notebooks and articles explaining some interesting techniques that you can use with TensorFlow and Keras.

Improve TensorFlow performance with Graps Mode.

Using the cats_vs_dogs dataset we are going to explore hot to use the Graph Mode and how it can improve the performance in our Notebooks.

Multiple Outputs with Keras Functional API.

Using the functional API and a non sequential Model we are going to learn ho to create a model able to predict more than one variable. The Datased used is the Wine_quality and we will predict a Classification Variable to classify the wine, and a Regression variable to qualify the wine.

How to create a siamese network with Tensorflow.

we are going to build a Siamese Model with Tensorflow able to compare images and return a difference between them, identifying when the images are of the same type.

Using Multiple GPUs with TensorFlow on Kaggle.

Recently, Kaggle has introduced the possibility of using two GPUs in our notebooks. We are going to see how to use them in a simple way with TensorFlow.

The technique of using more than one GPU on a single machine is called MirroredStrategy. It is the one we are going to use in Kaggle. Or on our machine if we had more than one GPU.

By the moment, that's all!

If you want to continue your learning path, I recommend you my repository about GANs with 4 notebooks organized as an Introductory course to GenerativeAI.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published