{"id":427,"date":"2021-05-31T19:46:16","date_gmt":"2021-05-31T15:16:16","guid":{"rendered":"https:\/\/m-shaeri.ir\/blog\/?p=427"},"modified":"2024-12-20T09:59:02","modified_gmt":"2024-12-20T06:29:02","slug":"tensorflow-keras-train-in-python-predict-use-in-c-plus","status":"publish","type":"post","link":"https:\/\/mshaeri.com\/blog\/tensorflow-keras-train-in-python-predict-use-in-c-plus\/","title":{"rendered":"TensorFlow-Keras Model, Train in Python, Use in C++"},"content":{"rendered":"\n

Despite many advantages of Python <\/strong>such as simple syntax, easy-to-read code and large number of available 3rd-party packages, when it comes to embedded systems and real-time applications, C++<\/strong> is undoubtedly one of the most efficient and powerful programming languages that dominates the realm of these systems. <\/p>\n\n\n\n

\"Embedding<\/a>
Top 10 languages that used in embedded projects (Reference<\/a>)<\/figcaption><\/figure>\n\n\n\n

However, complex syntax of C++<\/strong> makes it difficult to implement high level algorithm such as Machine Learning<\/strong> algorithms. For example building a small CNN network <\/strong>and it’s training pipeline can go so complex that no developer prefers doing such task the in C++<\/strong> from the scratch.<\/p>\n\n\n\n

In this post, we’ll learn how to transfer and use CNN models<\/strong> that is trained in Python <\/strong>with Keras <\/strong>to C++<\/strong> using a light weight header-only library named frugally-deep<\/strong>. I found no good tutorial on the web teaching this library and its usage in details. So, here we are going to explore it step by step.<\/p>\n\n\n\n

First we train a simple CNN network in Python on a 2000 Cat & Dog image dataset. No matter how much accuracy we achieve, we just save the model to use it in the c++ code : <\/p>\n\n\n\n

\n
\n
\n
\n \n
\n
\n \n
\n\n \n
\n\n