Skip to content
Building an image classifier in TF2
Jupyter Notebook
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
LICENSE
Practical_intro_to_TF2.ipynb
README.md
finetune_metrics.png
inception_model_with_shape_info.png
pretrained_losses_accuracies.png
sunflower.jpg
tensorboard_screenshot.png

README.md

Practical Introduction to TensorFlow 2.0

This repository introduces the new Tensorflow 2.0 in a practical way by building an image classifier which classifies 5 classes of flowers. It covers the following:

  1. Downloading and preprocessing data using TensorFlow Datasets
  • Checking out available datasets and their features
  • Downloading the dataset (tfds.load()))
  • Pre-processing the dataset
  • Visualizing the dataset
  1. Building and training an image classifier model using Keras high level API
  • Building a simple CNN in Keras
  • Visualising the model
  • Compiling and training the model
  • Training the model using data augmentation
  • Using TensorBoard inside notebooks
  1. Downloading and fine-tuning InceptionV3 pre-trained model
  • Downloading pre-trained model
  • Adding classification head
  • Training the classification head
  • Fine tuning the model
  1. Serving the trained model using TensorFlow Serving
  • Tensorflow Serving installation
  • Starting TensorFlow Serving
  • Making REST requests
  • Parsing the response

Read the Medium article explaining the above concepts in detail here.

You can’t perform that action at this time.