Student Symposium Poster for Model GUI Project:

Link

This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.

Allows Inception, ResNet, VGG, and LeNet models to be used with custom datasets without the need for using custom scripts tailored to each model. Helps non-experts with learning how to train and classify using existing pre-trained models.

As of the current TensorFlow version, r1.11, the LeNet model is no longer functional, as are the VGG Models. Currently working on a fix for these two models as well as implementing new types of classification (NLP, etc).

• Install tensorflow, matplotlib, numpy, and pygubu
• If .tfrecords files (train-00000-of-00001 and validation-00000-of-00001) have not been generated, use:

  ./run_build_image_data.sh

• Open GUI using:

  python model_wrapper_gui.py