Intended for uploading model to SSkin's GCP environment.
- Load trained model from kaggle
- Add base64 preprocessing layer for receiving input in base64 string format
- Upload it to your Google Cloud Storage
Before using this, the following requirements should satisfy:
- A kaggle.json, authorized to access the corresponding kaggle output URL. Must be placed on default directory of Google Colabs.
- A GCP service account's service key (.json), authorized to upload object to a corresponding project's bucket.
- Fork this repository
- Open the forked repository in google colabs
- Prepare kaggle.json and GCP SA key and place them accordingly
- Change the variables in source code to ur needs (like project name, bucket, and path)
- Run the whole thing
See LICENSE