Skip to content

CS 5542 BigData Lab Report #08

Jump to bottom
Amy Lin edited this page Mar 29, 2017 · 1 revision

TensorFlow PROGRAMMING - Softmax + Gradient Descent Algorithm

[ QUESTION ]

Implement SoftMax Classification for Image Dataset that is not covered in class. Report Accuracy.

Visulizations(Tensor Board): Training and Testing.

[ IMPLEMENTATION ]

  • Import CIFAR-10 data using data_helpers.py to unpack.
  • Connect tensorflow and python.
  • Input Place Holders for labels and bias.
  • Set weights and bias.
  • Run Softmax function.
  • Create different layers to add them together.
  • Cross Entropy calculating.
  • Train using gradient descent.
  • Save summaries for visualization.
  • Merge all summaries into one op.
  • FileWriter to graph in tensorboard.
  • Train data for 1000 steps.
  • Save the snapshot of the trained model for testing.
  • Restore the trained model for testing.
  • Calculate the accuracy of the model & plot the graph on tensor board.

< DATASET >

The CIFAR-10 DATASET

[ RESULTS ]

  • Cross Entropy Scalars



  • Bias Distribution



  • Bias Histogram



  • Cross Distribution



  • Cross Histogram



  • Weight Distribution



  • Wieght Histogram



  • Max Weight Histogram



  • SoftMax Graph



  • SoftMax Accuracy



Google Cardboard

[ QUESTION ]

Develop a Cardboard App that is relevant to your own project.

User input event handling.

! I haven't quite figured this one out yet!... : S

Clone this wiki locally