How is the weight generated when migrating and learning mobilenetv2 network in edge impulse?

[mengmeng]

How is the weight generated when migrating and learning mobilenetv2 network in edge impulse?

[aurel]

Hi @mengmeng,

You can read more about our transfer learning mechanism here:

Questions about Transfer learning Mechanism

In our transfer learning scheme we run two training cycles. The main one trains the final layer of the model from scratch with a high learning rate. The “fine tuning” cycle then uses a smaller learning rate to update a subset of the earlier layers. All of this is configurable in the script via FINE_TUNE_EPOCHS and FINE_TUNE_PERCENTAGE. Note that right now transfer learning doesn’t work in an iPython Notebook since we don’t provide the model weights. I’m working on a fix for this currently and i…

Aurelien

[mengmeng]

Thank you for your reply and look forward to your work！