Skip to content
No description, website, or topics provided.
Python Shell
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
DataCollection
Models/CNN
Results
Server
jp_Results
README.md
data_preprocess.sh
datacleaned.json
git_prep.sh
load_data.py
run.py
run_java_python.py
training_options.py
view_augmented.py

README.md

Video Code Tagging

This repository contains work for two projects. The first is interested in identifying the presence of Java code in software engineering video tutorials. The second is interested in predicting Java versus Python code in image frames.

⇒ See Labeling Standards for info on how we label data
⇒ See DataCollection for how we aquire data
⇒ See Models for the networks we use to label images
⇒ See Results for results on identifying Java code in videos
⇒ See Java Python Results for results of discriminating between Java and Python code in videos
⇒ See Server how our tagging tool works on the backend

Steps

  1. Acquire links of videos
  • Run bash DataCollection/grab.sh to pull all the videos and split them into frames
  1. Label data
  • Ex: Image that contains Java code
    • path/to/img,1,0,0,0
  1. Resize data set
  • Run mv data_preprocess.sh path/where/dataset/is
  • Run bash data_preprocess.sh resize all images in dataset
  1. Train your model
  • Run python run.py
    • This loads training set with labels and begins training
You can’t perform that action at this time.