Skip to content

A series of programs written in Python using Weightling App data to train a Keras deep learning model exported to CoreML for the purpose of workout type prediction based on the exercises performed

Notifications You must be signed in to change notification settings

WeightliftingApp/WeightliftingApp-ML

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

WeightliftingApp ML Model

A series of programs written in Python using Weightling App data to train a Keras deep learning model for the purpose of workout type prediction based on the exercises performed.

Body split training input example:

"2 Abs / Core#3 Biceps#2 Triceps" becomes [ 2 0 3 0 0 0 0 0 2 ] :

Abs / Core back Biceps cardio chest legs Olympic shoulders Triceps
2 0 3 0 0 0 0 0 2

Categorical expected output example:

"arms" becomes [ 0 1 0 0 0 0 0 0 0 0 0 0 ] :

abs arms back cardio chest legs shoulders push pull chestBack chestBiceps fullBody
0 1 0 0 0 0 0 0 0 0 0 0
      ||||||||
      ||||||||
      ||||||||
  Processed by model
    \\\\\\//////
      \\\\////
        \\//

Categorical actual output example:

[ .003 .981 .001 .002 .000 .000 .000 .000 .000 .000 .012 .001 ] becomes "arms" :

abs arms back cardio chest legs shoulders push pull chestBack chestBiceps fullBody
.003 .981 .001 .002 .000 .000 .000 .000 .000 .000 .012 .001

Current accuracy (744 training points @ 100,000 epochs, categorical crossentropy):

  • Training data: 100.0% 👍
  • Test data: 97.0% 👍

Current model

Model Image

About

A series of programs written in Python using Weightling App data to train a Keras deep learning model exported to CoreML for the purpose of workout type prediction based on the exercises performed

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages