Skip to content

MCRen88/npi_bubblesort

Repository files navigation

npi_bubblesort

Nerual program interpreters of bubblesort

Authors(List by name)

  • Junling Chen
  • Xingye Xu

This project is based on these main resources:

  1. DeepMind's ICLR 2016 best award: Neural Programmer-Interpreters.
  2. The great Addition development of the DeepMind ideas that @mokemokechicken did in his repo: https://github.com/mokemokechicken/keras_npi

requirement

  • Python3

setup

pip install -r requirements.txt

tips: Keras use theano as backend. If you use tensorflow, it maybe many problems.

create training dataset

create training dataset

sh src/run_create_bubblesort_data.sh

create training dataset with showing steps on terminal

DEBUG=1 sh src/run_create_bubblesort_data.sh

training model

Create New Model (-> remove old model if exists and then create new model)

NEW_MODEL=1 sh src/run_train_bubblesort_model.sh

Training Existing Model (-> if a model exists, use the model)

sh src/run_train_bubblesort_model.sh

test model

check the model accuracy

sh src/run_test_bubblesort_model.sh

check the model accuracy with showing steps on terminal

DEBUG=1 sh src/run_test_bubblesort_model.sh

About

Nerual program interpreters of bubblesort

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published