The people behind TensorFlow soon took note of this, and adopted many of PyTorch’s most popular features in TensorFlow 2.0. A good rule of thumb is that you can do anything that PyTorch does in ...

PyTorch recreates the graph on the fly at each iteration step. In contrast, TensorFlow by default creates a single data flow graph, optimizes the graph code for performance, and then trains the model.

No longer the upstart nipping at TensorFlow’s heels, PyTorch is a major force in the deep learning world today, perhaps primarily for research, but also in production applications more and more.

TensorFlow is your ally for scalability and production. PyTorch is your friend for research flexibility and ease of use. The choice depends on your project needs, expertise, and long-term goals.

Available today, PyTorch 1.3 comes with the ability to quantize a model for inference on to either server or mobile devices. Quantization is a way to perform computation at reduced precision.