What is the name of Google's open-source machine learning framework?

Google’s open-source machine learning framework is TensorFlow.

TensorFlow is an open-source software framework developed by the Google Brain team for numerical computation and machine learning. Its name refers to multidimensional arrays, commonly called tensors, flowing through a computational graph. This structure lets developers describe data-processing and model-training operations that can run efficiently on CPUs, GPUs, and specialized hardware.

Google released TensorFlow as open-source software in November 2015. It grew from an internal Google system called DistBelief and became widely used for tasks such as image recognition, speech processing, natural-language systems, and predictive modeling. TensorFlow also includes tools for building, training, deploying, and monitoring machine-learning models.

Caffe, Keras, and PyTorch are common sources of confusion. Caffe and PyTorch are separate machine-learning frameworks, while Keras is a high-level deep-learning API that can work with TensorFlow as a backend. TensorFlow is therefore the canonical answer to this wording, even though modern developers may use several of these tools together.