Unload the model's weights from GPUs. Optional, but hints that the user is done using the model for now; allows Tinker servers to free resources.
Unload the model weights and ends the user's session.
post/api/v1/unload_model
Request body
Example request
{
"model_id": "123e4567-e89b-12d3-a456-426614174000"
}Response
API future for unloading the model
Example response
{
"request_id": "123e4567-e89b-12d3-a456-426614174000",
"model_id": "123e4567-e89b-12d3-a456-426614174000"
}