Open Source Summit + ELC North America 2020

The current machine learning hardware landscape is dominated by Nvidia and its closed-source CUDA ecosystem. Although AMD's ROCm software is a growing rival, Nvidia is still the market leader by a wide margin. Open Source machine learning libraries such as TensorFlow, PyTorch and mlpack rely on this closed stack to accelerate learning and inference, locking developers in.

Can we give freedom of choice back to machine learning developers?

The new Panfrost open source driver for Mali GPUs is solving this problem on the Edge by enabling a fast and efficient machine learning stack running pure open source.

In this talk, we will walk through the process of building a computer vision pipeline on top of a completely open-source inference stack: open-source GPU driver, machine learning framework and machine learning models. We will also share what we have learned about optimizing these models to run fast on resource-constrained hardware such as the Rockchip RK3399.