AI-Assisted Workflows Reference

AI-Assisted Workflows Reference#

This section describes each AI-assisted workflow that ships with Vitis AI.

The workflows are optional. Each task is fully documented for manual execution, and the manual procedure is linked from every workflow page. If you do not use an AI coding assistant, follow those links instead. To set the workflows up, see Workflow Setup.

Choosing a workflow#

If you want to

Workflow

Manual procedure

Implement a custom operator that can extend what the NPU can run and/or optimize how it runs.

Custom operator workflow

Custom operators

Quantize a model, including per-layer mixed precision

Model quantization workflow

Model Quantization , Mixed Precision Compilation

Increase NPU offload, reduce inference time, or get a model to compile

Compiler options workflow

Vitis AI EP Options

Board access#

The workflows steer themselves using measurements taken on hardware. Without board access they fall back on simulation and host-side compile results, which are not a one-to-one match for the hardware. They still run, but expect numeric mismatches, untuned performance, and weaker convergence. Each workflow page describes what it can still do without a board.

The workflows reach the board over SSH, using either a password or a key. Set the following before launching your assistant:

export BOARDHOST=172.16.0.1        # your board's IP address
export BOARD_USER=amd-edf

# Authenticate with a password
export BOARD_PASSWORD=<password>

# Or with an SSH key, instead of a password
# export BOARD_KEY=/mnt/user/ssh_key_file

Caution

BOARD_PASSWORD is a credential. Keep it out of scripts and version control, and prefer a permission-restricted file over a plain shell export. See the guidance in Configure model access.

To persist these across container restarts, append them to $HOME/.bashrc, which is $WORKDIR/homedir/.bashrc on the host. See Example: Claude Code in the Vitis AI Docker container.

If you prefer not to set them, supply the board details in the prompt instead:

Run the model ./model.onnx on my board using configuration
./vitisai_config.json. The board's IP address is 172.16.0.1, the user is
amd-edf, the password is mypassword.

Boot the board and confirm it is reachable over SSH before you start a session. See the OSPI and SD card boot flow documentation OSPI and SD Card Boot Flow for board preparation and bootup.

Contents