Skip to content

Latest commit

 

History

History
14 lines (14 loc) · 515 Bytes

README.md

File metadata and controls

14 lines (14 loc) · 515 Bytes
Raw

Digit-Recognizer

Digit Recognition on MNIST Data.
If you want to use the trained weights for your own testing, initialize all the variables first and then restore the saved model. It will overwrite the pre-weights and you will have the trained weights for your own testing purpose.

Libraries Used:

  • Numpy
  • Tensorflow

Network:

  • Multilayer Perceptron

Optimizer:

  • AdamOptimizer

Hidden Layers:

  • 2

No. of nodes in both Hidden Layers:

  • 256