Start a Training and Poll Until Complete
A 3-step recipe with Replicate. We can help you get access.
Recipe · 3 steps
Start a Training and Poll Until Complete
with Replicate · Artificial Intelligence Management
Start a training from a base version and poll it until it finishes.
- Confirm the base model version exists and is trainable before starting a training run.
- Start a training of the base version, writing the resulting version to the destination model. The training is returned in a starting state.
- Retrieve the training state, repeating via the retry branch until the training reaches a terminal status.
What you'll need to have ready
- apiToken: Replicate API token used as a Bearer credential.
- modelOwner: The owner of the base model being trained.
- modelName: The name of the base model being trained.
- versionId: The ID of the base model version to train.
- destination: The destination model in the form {owner}/{name} to push the trained version to.
- input: Inputs to the Cog model's train() function.
- webhook: Optional HTTPS URL to receive a webhook when the training completes.
For your developer
Arazzo workflow replicate:train-model-and-poll, 3 operations: models.versions.get → trainings.create → trainings.get. Specs on APIs.io.