In our Introduction to Neural Networks you could get familiar with the basics of machine learning. But you shouldn’t stop exploring that huge field of knowledge. There are so many interesting things and we want to present to you one of them. TensorStream ML framework that supports pure Ruby.
TensorStream is an opensource framework for machine learning for Ruby, its goal is to allow machine learning models to be built, so you can run them in various hardware like GPUs and CPUs. TensorStream is made to support various backends with a Pure Ruby and OpenCL implementation.
The main feature of TensorStream is Tensors. They scalar, single and multidimensional arrays in a consistent manner. Tensors have properties that describe their shape and rank as well as their data type.
- The shape of a tensor describes its structure or describes the dimensions of the array.
- TensorStream supports all of the basic math operations you would expect, only beefed up to work with tensors.
- Tensor sizes don’t have to be the same, you can, in some instances use a different but compatible size in order to perform an operation.
- There are special types of tensors that are frequently used in building a model in order to serve as placeholders as well as to store data that can be used in succeeding sessions – variables.
- Graphs hold the entire model data structure, each operation defined is stored in a graph which is later used during runtime to perform operations.
- TensorStream has been designed from the ground up to support multiple execution backends.
Feel free to check more information here.
For you to know, recently we made a small list of ML and DL best libraries.
It will be worthwhile to read about top 10 automation testing tools to take advantage of these trends. Stay tuned!