Converting a model from Pytorch to Tensorflow: Guide to ONNX

Open Neural Network Exchange (ONNX) is a powerful and open format built to represent machine learning models. The final outcome of training any machine learning or deep learning algorithm is a model file that represents the mapping of input data to output predictions in an efficient manner. These models are stored in different file formats depending on the framework they were created in .pkl Therein lies the problem, you can't take a model created and trained in one framework and use it or deploy it in a different framework. The intent behind ONNX is to be like the "USB standard" of the machine learning world.