I have compressed the saved model as an h5 file and compressed it using CoreML. I have now created a user-interactive, friendly-to-use application so that you can upload a fundal image to get an accurate diagnosis.
gajanm/DiaDetectPhone
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|