CODEWITHSUNDEEP
tensorflowgoogle-cloud-platformgoogle-cloud-visiongoogle-cloud-automlautoml
Thomas Dorsey
Thomas Dorsey

Reputation: 15

How to train for edge tpu on google cloud vision with object detection?

When I try to export a model for the Google Coral, the option that should exist doesn't exist. As per this link, there should be an option to export for the Coral, but for me only TF Lite, Tensorflow.js and Container exist as options.

Upvotes: 0

Views: 300

Answers (1)

aemon4
aemon4

Reputation: 1102

The link you passed to the documentation [1] refers to image classification models (single label or multi-label classification). In the case of object detection exporting to EdgeTPU [2] is not supported and there is no current workaround.

Thus, it is normal that you do not see the option.

[1] https://cloud.google.com/vision/automl/docs/export-edge#edge-tpu

[2] https://cloud.google.com/vision/automl/object-detection/docs/export-edge

Upvotes: 1

Related Questions