Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.osmosis.ai/llms.txt

Use this file to discover all available pages before exploring further.

Deployments let you serve a trained LoRA checkpoint after a training run finishes. You can deploy from the platform UI or from the CLI while working inside your workspace directory.

​
Deploy a Checkpoint

After a training run finishes, inspect it to find available checkpoints: 
osmosis train info <run-name>
Deploy a checkpoint by checkpoint name or ID: 
osmosis deploy <checkpoint-name>
If you run osmosis deploy without an argument in an interactive terminal, the CLI prompts you to choose a training run and checkpoint.

​
Inspect Deployments

List deployments: 
osmosis deployment list
Show deployment details: 
osmosis deployment info <checkpoint-name>
Deployment details include the checkpoint, status, base model, step, creator, and creation time.

​
Undeploy

To deactivate a deployed checkpoint: 
osmosis undeploy <checkpoint-name>
The checkpoint remains available in the training run history; undeploy only transitions the serving deployment to inactive.

​
Requirements

  • Run deployment commands from the workspace directory so the CLI can resolve the connected workspace from Git origin.
  • The checkpoint must belong to a training run in the same workspace.
  • GitHub setup must be healthy before training can produce new checkpoints.

​
Next Steps

Training Runs

Submit training runs and inspect checkpoints.

Command Reference

Review deployment commands and options.