How-To · Access Control

How-To: Assign Token to Workload

This guide shows how to give an existing User Access Token access to a model: from the model itself, from the Tokens page, or with the Python SDK. To create a new token for a model instead, see Create User Access Token.

Before you start

Two ways in the Console

From the model, the Add token list shows unassigned tokens; selecting one assigns it to that model. From the Tokens page, Edit token lets you choose the needed model directly, or move an assigned token straight to another model.

Not while a worker runs

A token's model can't be changed or removed while a worker is running on its current model. Stop the worker first.

New model, new key

A token moved to a different model needs a new evaluation key before it can be used for queries.

Generate a new evaluation key after a move

Evaluation keys are tied to a token and its model. After you assign a token to a different model, generate and upload a new key for it. See Upload Evaluation Key.

Assign a token

Console: from the model

Use this to connect an unassigned token to a model you're looking at.

console.lattica.ai/models1 / 6

1

Step 1 of 6:

Open Models

In the Console, select Models in the left menu.

Console: from the Tokens page

Use this to assign any token, or to move an assigned token straight to another model.

console.lattica.ai/tokens1 / 5

1

Step 1 of 5:

Open Tokens

In the Console, select Tokens in the left menu and find the token.

Python SDK

Assign a token with studio.tokens.assign. It returns the server's message, plus a warning when there is one, for example when the token isn't Active.

assign_token.py

import os

from lattica_studio import LatticaStudio

studio = LatticaStudio(os.environ["LATTICA_LICENSE_KEY"])

model = studio.models.get_by_name("DermAI Spotter")
token = next(t for t in studio.tokens.list() if t.name == "partner-trial")

# To move a token that is assigned elsewhere, remove that assignment first:
# if token.model_id:
#     studio.tokens.unassign(token.id, token.model_id)

result = studio.tokens.assign(token.id, model.id)
print(result["message"])
if result.get("warning"):
    print("Warning:", result["warning"])

After assigning, the token holder generates and uploads an evaluation key for the model before querying. See Upload Evaluation Key.