carlgira / oci-image-classification

OCI image classification using data labeling and vision service

OCI Image Classification

The idea with this repository is to create a custom dataset using the datalabeling and vision service of Oracle Cloud.

I wrote a blog explaining how to use this repository https://medium.com/@carlgira/ai-vision-model-on-oci-for-image-classification-90387524d69d

I'll be using the landmaks dataset v2 https://github.com/cvdfoundation/google-landmark.

Initial Setup

1. Follow the instructions to configure oci cli and to create API keys to authenticate against OCI that are in this blog https://medium.com/@carlgira/install-oci-cli-and-configure-a-default-profile-802cc61abd4f

2. To start the notebook create a conda enviroment using the next commands.

    conda env create --name oci-custom-dataset-env --file environment.yml
    conda activate oci-custom-dataset-env

3. Once the enviroment is activated start jupyter.

    jupyter-notebook