Skip to content
Website for oxykodit, including online Deep Learning demo with Tufa image recognition
Branch: master
Clone or download
Latest commit 331c3d0 Jan 21, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
ml_model ML model URL as parameter Jan 7, 2019
static aesthetic changes Jan 15, 2019
templates aesthetic changes Jan 15, 2019
.gitignore cleanup Jan 2, 2019
README.md Update README.md Jan 14, 2019
app.py randomly showing messages while training the model Jan 14, 2019
ml_model.py from 13s to 8s Jan 21, 2019
requirements.txt fixes to generate in-mem predictions Jan 2, 2019

README.md

Python & Flask code to run Tufa example @ OxyKodit website

app.py contains the Flask framework that parses GET/POST information, keeps track of the image annotations, trains the Machine Learning model and runs the predictions.

ml_model.py contains the Python Machine Learning code that takes the pretrained neural net and retrains the last layers with the annotated Tufa examples. This file is based on code from the Tensorflow github.

ml_model contains the pre-trained models cached from https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/feature_vector/2 and https://tfhub.dev/google/imagenet/mobilenet_v2_035_128/feature_vector/2

templates contains the HTML files, including javascript, for rendering the interactive grid page.

static contains the image JPG files.

You can’t perform that action at this time.