question

GGG9 avatar image
0 Votes"
GGG9 asked GiftA-MSFT edited

Model training loss function in Custom Vision


What loss function does Custom Vision use to train a model for classification?



azure-custom-vision
· 1
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

Hi, we are inquiring on this information, we'll get back to you as soon as possible.

0 Votes 0 ·

1 Answer

GiftA-MSFT avatar image
0 Votes"
GiftA-MSFT answered GiftA-MSFT edited

Hi, here's the feedback I received:

For image classification, we use cross-entropy loss.
For object detection, we use bounding box regression loss for bounding box accuracy and cross-entropy loss for object category.



5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.