Skip to content

Publish

Once you've completed the model training and saved the results, the next step is to publish your model. This section guides you through using the nimble-cli to make your model available on the network.

Listing Models Ready for Publication

Before you can publish a model, you need to identify which ones are fully trained and ready. Models still in training are not eligible for publication.

python
nimble-cli list models

This command will display a list of models that are ready to be published.

Committing Your Model

Choose a model to publish. You will need to provide the model's ID and your signature to commit it to the network.

python
nimble-cli publish 
--id
--storage
--price_tag
--validator_type
--signature

Validation and Consensus

Once submitted, your model will undergo a validation process where a model quality consensus algorithm assesses whether the model was properly trained and holds meaningful insights. This step ensures the integrity and reliability of models on the Nimble Network.

Final Note

After your model is published, no further changes can be made to it. If you need to adjust the model or wish to build upon it, you must create a new trainer instance and go through the publication process again. This ensures that each model version remains distinct and traceable.