Skip to content

MobileNet v2 #8

@Riyya-HI

Description

@Riyya-HI

I'm trying to retrain mobilenet v2 such that I already have a pre-trained model (on coco) with weights and I just want to train the last few layers with new images with objects from coco and the model just fine-tunes to that to increase accuracy without losing out on any pre-existing knowledge of all objects on coco. Basically just fine tune to new images. Not customization. How can I do this?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions