89
edits
Changes
Added training history
</syntaxhighlight>The training was done in the cloud using a AWS EC2 server setted up ad hoc.
The dataset was created collecting 240 images of 6 different fruits. 3/4 of the images were used for the training and the rest was used for test /validation purpose. Training the model with a greater number of images would have led to a better accuracy but it wouldn't have changed the inference time.
Several measures were taken to counter the high overfitting tendency due to the small amount of images. For instance, new images were synthesized from the existing ones to simulate a larger dataset, as shown below:
[[File:Drawio200% augmentation.png|none|thumb|400x400px|New images synthesized from an existing one. Original image by tookapic from Pixabay.com]]
The following plots show the training history:
[[File:Keras loss history.png|none|thumb|400x400px|Variation of the loss (blue) and the validation loss (orange) through the epochs during training]]
[[File:Keras acc history.png|none|thumb|400x400px|Variation of the accuracy (blue) and the validation accuracy (orange) through the epochs during training]]
==Articles in this series==
