/Digit-Recognition-with-TensorFlow

A TensorFlow project for training and testing a feedforward neural network on the MNIST dataset, with data organized in a folder and a JSON label file.

Primary LanguageJupyter Notebook

Digit-Recognition-with-TensorFlow

This TensorFlow-based project focuses on digit recognition using the MNIST dataset. It includes a Jupyter Notebook for model training and evaluation.

Contents

  • Digit_Recognition.ipynb: Jupyter Notebook for model training and evaluation.
  • Data/: Folder containing the dataset.
  • labels.json: JSON file with label information.
  • Digit_Recognition.zip: Zip archive containing the dataset and JSON file.
  • mnist_model.h5: Trained model file for digit recognition.

Usage

  1. Open the Digit_Recognition.ipynb Jupyter Notebook to work with the project.
  2. The dataset is stored in the Data/ folder, and labels are available in labels.json.
  3. To access the full dataset, download and extract Digit_Recognition.zip.

Performance Plots

  • accuracy.png: The plot displays the model's accuracy during training.
  • loss.png: The plot shows the model's loss over the training epochs.