Convolutional Neural Network are current state of the art for image classification. Let's look how we can build one. 👍 You can find here examples of high level
and low level
Tensorflow implementations.
The Convolutional Neural Network recognizes a dug toy from our custom dataset. Yes, you can change dataset and train your own CNN.
- How to define Tensorflow graph
- How to train the network with custom dataset
- How to freeze the model to
.tflite
- This repo is part of our blog post eMan Tensorflow
- If you're interested in implementing frozen model to an Android app via Tensorflow Lite, check our Simple Android Tensorflow Lite example
- If you find this repo useful, please consider ★ starring it