TFDEV

A starter for tensorflow, I try to let it contain the convenience of keras and the flexibility of tensorflow.

Functionality

• Integrate Dataset API
• Bultin Layer
  • Batch Normalization
  • Conv2d Layer
  • Dense Layer
  • Dropout
• Regularization
• Callback
  • Early Stopping
  • Tensor Board
  • Check Point
• Session Management
• Re-Build Model

Future Work

• Mnist, Cifar Dataset
• RNN
• Inject
  • Tensor Board

Getting Started

1. get mnsit data, see and execute demo/download.py (make sure all of generated files in demo/)
2. define your model, see demo/instances.py
3. setup dataset and train model, see demo/train.py