{"id":10118,"date":"2022-10-04T15:54:39","date_gmt":"2022-10-04T19:54:39","guid":{"rendered":"https:\/\/blogs.mathworks.com\/deep-learning\/?p=10118"},"modified":"2024-04-16T16:49:33","modified_gmt":"2024-04-16T20:49:33","slug":"whats-new-in-interoperability-with-tensorflow-and-pytorch","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/deep-learning\/2022\/10\/04\/whats-new-in-interoperability-with-tensorflow-and-pytorch\/","title":{"rendered":"What’s New in Interoperability with TensorFlow and PyTorch"},"content":{"rendered":"For deep learning, MATLAB allows users to create and train models in MATLAB or leverage models trained in open source via model conversion. Prior to MATLAB R2022b, support for model conversion included: import from and export to ONNX\u2122, and import from TensorFlow\u2122. We are excited to share that as of MATLAB R2022b, users can now export models to TensorFlow<\/strong> as Python\u00ae code and can import models from PyTorch\u00ae<\/strong> (starting with support for image classification).\r\n
<\/h6>\r\n \r\n

Export to TensorFlow<\/strong><\/p>\r\n

\"\"<\/p>\r\n\r\n

<\/h6>\r\nThe support package Deep Learning Toolbox Converter for TensorFlow Models<\/a> just added the capability to export from MATLAB to TensorFlow, by using the\u00a0exportNetworkToTensorFlow<\/span><\/a> function.\r\n
<\/h6>\r\nThere are many reasons to be excited about the new exportNetworkToTensorFlow<\/span> function:\r\n
<\/h6>\r\n\r\n\r\n\r\n
\r\n
    \r\n \t
  1. You can export deep learning networks and layer graphs (e.g., convolutional, LSTM) directly to TensorFlow.<\/li>\r\n \t
  2. It\u2019s easy to export a network, load it as a TensorFlow model, and use the exported model for prediction. For a relevant example, see Export Network to TensorFlow and Classify Image<\/a>.<\/li>\r\n \t
  3. It\u2019s easy to export an untrained layer graph, load it as a TensorFlow model, and train the exported model. For a relevant example, see Export Untrained Layer Graph to TensorFlow<\/a>.<\/li>\r\n \t
  4. You can save the exported model in any standard TensorFlow format, such as SavedModel or HDF5 format, and share it with your colleagues who work in TensorFlow.<\/li>\r\n \t
  5. Many MATLAB layers can be converted to TensorFlow layers. For a full list of the layers, see Layers Supported for Export to TensorFlow<\/a>.<\/li>\r\n<\/ol>\r\n<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>\r\n
    <\/h6>\r\nHere I am showing the basic workflow of how to export a deep learning network to TensorFlow, load it as a TensorFlow model, and save it in SavedModel format.\r\n
    <\/h6>\r\n\r\n\r\n\r\n
    MATLAB Code:<\/span>\r\n
    <\/h6>\r\nLoad a pretrained network. The\u00a0Pretrained Deep Neural Networks<\/a> documentation page shows you all options of how to get a pretrained network. You can alternatively create your own network.\r\n
    <\/h6>\r\n
    net = darknet19;<\/pre>\r\n
    <\/h6>\r\nExport the network net<\/span> to TensorFlow. The exportNetworkToTensorFlow<\/span> function saves the TensorFlow model in the Python package DarkNet19.\r\n
    <\/h6>\r\n
     exportNetworkToTensorFlow(net,\"DarkNet19\")<\/pre>\r\n<\/td>\r\n<\/tr>\r\n<\/tbody>\r\n<\/table>\r\n
    <\/h6>\r\nThe\u00a0DarkNet19\u00a0package contains four files:\r\n