Bring your own LLM/DL workloads requires analytic gpu computer clusters. These are configured with NVIDIA GPU accelerated containers that are designed for model inference jobs using deep learning and large language models. To set up the environment using SQL, see Set Up Lake for LLM/DL Using SQL.I
- Connect to the environment with admin credentials..
- Select Compute groups .
- Create a group ().
- Select Analytic GPU as the group type.
- Select the new group and Create a profile.
- Go to Manage access then Users.To access Users, you'll need the root (DBC) password.
- Create new or edit existing users to grant access to the new analytic GPU compute group.
- Grant the user or user groups permission to execute the APPLY operator.
GRANT EXECUTE ON FUNCTION TD_SYSFNLIB.APPLY TO username; - Grant permission to create and manage the User Environment using Open Analytics Framework APIs.
GRANT TD_DATA_SCIENTIST to username;