Install & Manage¶
Download, update, and remove models.
Install a Single Model¶
import modeldock as md
md.install("llama3")
md.install("llama3", backend="ollama") # explicit backend
Install by Category¶
Bulk install recommended models for a specific use case:
Check What's Installed¶
Update a Model¶
update() removes then re-downloads — it's destructive and requires confirmation:
Destructive Operation
update() deletes the model and re-pulls it. Always use confirm=True. Cloud/subscription models are rejected.
Remove a Model¶
Verify Integrity¶
Runtime Status¶
Check if the runtime is available and which device models execute on:
device is Device.GPU, Device.CPU, or Device.UNKNOWN.
Next Steps¶
- Load & Run — load models and run completions
- Cache Management — manage cached artifacts