Tensorflow-based demo for recognition of hand-written digits. Training is performed on the MNIST dataset. The demo is based on static graph construction.
mnist_slim.ipynb -> Build, train and test the model
Tensorflow-based demo for recognition of hand-written digits. Training is performed on the MNIST dataset. The demo is based on TensorFlow Eager Executeion (TF >= 1.7)
mnist_eager.ipynb -> Build, train and test the model
Tensorflow-based demo for binary image classification [Cat | Dog]. Training is performed on the ImageNet subset. The demo is based on static graph construction.
Tensorflow_train_model.ipynb -> Build and train the model
Tensorflow_freeze_graph.ipynb -> Freeze Tensorflow graph (prepare for the inference)
Tensorflow_inference.ipynb -> Inference