Skip to content

Latest commit

 

History

History
39 lines (29 loc) · 839 Bytes

ImageClassification-Notes.md

File metadata and controls

39 lines (29 loc) · 839 Bytes

Training in Keras and Load in JS O/P difference between Python and JS .predict vs execute (LoadFrozenModel)

Which layer of model to use in Transfer learning

Training in browser? Adding text features to transfer learning Saving model in user's browser/local storage Running an image through multipe models having multiple classifications for an image argMax and softMax

Browser choosing files choosing model locations other param image size

Providing Feedback on the image Tags

Train every month using, last 12 months images Samples should have same straification as overall image month,label, page? location, user

Publishing trained model to server/cloud storage

Scehduling Training

Training on server Training mobilenet

model rotation scribble receipt mode