tao-launch-workflow

To launch a TAO training workflow on SLURM, you select the SLURM platform during the launch process, provide your cluster credentials, and run preflight checks. The workflow automates platform selection and environment verification before the job starts.

Yes, you can run TAO evaluation and inference tasks on Kubernetes. The workflow launch supports multiple execution platforms including Kubernetes, local Docker, Brev, and SLURM, handling platform selection and credential management for your tasks.

Preflight checks for TAO model training verify image resolution, compute shape, and dependencies. Running these checks before launching the workflow ensures the execution environment is correctly configured and prevents job failures.

Yes, you need platform and model-specific credentials to automate TAO workflows on Brev. The launch process handles credential management securely, requiring you to provide the appropriate credentials for your chosen platform before executing tasks.

The best way to streamline complex TAO export and training tasks is using a unified workflow launch. It automates platform selection, credentialing, and preflight checks, reducing manual setup overhead across different execution platforms.