amc-setup-calibration-stack

par nvidia

Lancer le microservice AutoMagicCalib et l'interface web à partir des images de version NGC via Docker Compose. Utiliser lorsque l'utilisateur dit 'déployer l'auto-calibration', 'lancer l'auto…

npx skills add https://github.com/nvidia/skills --skill amc-setup-calibration-stack

Skill: Launch AutoMagicCalib Release Containers

Set up the AutoMagicCalib microservice and UI from release containers: resolve an AMC checkout, authenticate to NGC, optionally download VGGT, configure Docker Compose, launch services, and verify readiness.

Prerequisites

  • Docker and Docker Compose installed
  • NVIDIA Docker Runtime configured (for GPU support)
  • auto-magic-calib repo on disk. Step 0b resolves the current repo, DeepStream tools/auto-magic-calib, DEEPSTREAM_REPO_ROOT, or ~/auto-magic-calib; otherwise it asks before cloning https://github.com/NVIDIA-AI-IOT/auto-magic-calib.
  • NGC account with access to NVIDIA container registry
  • Docker runnable without sudo; verify with docker ps before continuing.

Instructions

Step 0: Verify Docker Runs Without sudo

docker ps
  • If it succeeds → continue.
  • If it fails with "permission denied" → the user is not in the docker group. Ask the user to run:
    sudo usermod -aG docker $USER && newgrp docker
    
    Then ask the user to confirm docker ps works before continuing.

Agent note: If docker ps cannot be run from within the agent sandbox, ask the user to confirm it works (e.g. "Can you confirm docker ps runs without sudo?") before proceeding.

Step 0b: Resolve Repo Checkout

The skill needs AMC repo assets (compose/, sample data, and models/). Resolve an existing checkout first; ask before cloning into ~/auto-magic-calib.

REPO_URL="https://github.com/NVIDIA-AI-IOT/auto-magic-calib.git"
DEFAULT_CLONE_DIR="$HOME/auto-magic-calib"
CURRENT_GIT_ROOT="$(git rev-parse --show-toplevel 2>/dev/null || true)"

is_amc_checkout() {
  [ -n "$1" ] \
    && [ -f "$1/README.md" ] \
    && grep -q "AutoMagicCalib" "$1/README.md" 2>/dev/null \
    && [ -f "$1/compose/compose.yml" ] \
    && grep -q "auto-magic-calib-ms" "$1/compose/ms/compose.yml" 2>/dev/null \
    && grep -q "auto-magic-calib-ui" "$1/compose/ui/compose.yml" 2>/dev/null
}

REPO_ROOT=""
for candidate in \
  "$CURRENT_GIT_ROOT" \
  "${CURRENT_GIT_ROOT:+$CURRENT_GIT_ROOT/tools/auto-magic-calib}" \
  "${DEEPSTREAM_REPO_ROOT:+$DEEPSTREAM_REPO_ROOT/tools/auto-magic-calib}" \
  "$PWD/tools/auto-magic-calib" \
  "$DEFAULT_CLONE_DIR"; do
  if is_amc_checkout "$candidate"; then
    REPO_ROOT="$candidate"
    echo "✓ Using auto-magic-calib checkout: $REPO_ROOT"
    break
  fi
done

if [ -z "$REPO_ROOT" ]; then
  if [ -n "$CURRENT_GIT_ROOT" ] && [ -d "$CURRENT_GIT_ROOT/tools/auto-magic-calib" ]; then
    echo "Found $CURRENT_GIT_ROOT/tools/auto-magic-calib, but it is not an initialized AMC checkout."
    echo "If running from the DeepStream repository root:"
    echo "  git submodule update --init tools/auto-magic-calib"
  fi

  # Nothing usable on disk — STOP and ask the user for confirmation using the
  # host's question mechanism; if none is available, ask in chat and wait.
  # Do NOT clone silently from this block or clone over a tracked submodule path.
  echo "No usable auto-magic-calib checkout found. Ask the user for confirmation:"
  echo "  Clone $REPO_URL into $DEFAULT_CLONE_DIR? [y/N]"
  echo "On 'y' — run: git clone \"$REPO_URL\" \"$DEFAULT_CLONE_DIR\""
  exit 1
fi

cd "$REPO_ROOT"
export REPO_ROOT
echo "REPO_ROOT=$REPO_ROOT"

Agent note: never clone silently. Prefer initialized DeepStream tools/auto-magic-calib; do not clone over that submodule path. If it exists but is empty, ask the user to run git submodule update --init tools/auto-magic-calib. Honour an alternate AMC path if provided.

Step 0c: Install Python venv (New Systems Only)

On a fresh system, pip and python3-venv may not be available. Install them first:

# Create a venv for HuggingFace CLI (project-local preferred)
REPO_DIR="$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
HF_VENV="${REPO_DIR}/venv"
python3 -m venv "$HF_VENV" 2>/dev/null || {
  echo "ERROR: python3-venv not available." >&2
  echo "Install it manually: sudo apt install -y python3-venv python3-pip" >&2
  exit 1
}

# Install HuggingFace hub (needed for VGGT download)
"$HF_VENV/bin/pip" install --upgrade pip huggingface_hub

Note: Skip this step if a venv with hf already exists (check venv/bin/hf in the repo root or ~/venv/amc/bin/hf).

Step 1: Login to NGC

Ask the user for their NGC API key using the host's question mechanism; if none is available, ask in chat and wait. Then run:

echo "<NGC_API_KEY>" | docker login nvcr.io --username '$oauthtoken' --password-stdin
echo "✓ NGC authentication complete"

Step 2: Download VGGT Model (If Not Already Present)

export REPO_ROOT=$(git rev-parse --show-toplevel)
cd "$REPO_ROOT"

if [ -f "models/vggt/vggt_1B_commercial.pt" ]; then
  echo "✓ VGGT model already present"
else
  echo "✗ VGGT model not found"
  echo "Options:"
  echo "  1. Continue without VGGT (AMC only - sufficient for most use cases)"
  echo "  2. Download VGGT model (~4.7GB, requires HuggingFace account)"
fi

To download VGGT: ask the user to accept the license at https://huggingface.co/facebook/VGGT-1B-Commercial and provide a read token from https://huggingface.co/settings/tokens using the host's question mechanism. Pass it through HF_TOKEN so it is not exposed in ps output:

REPO_DIR="$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
cd "$REPO_DIR"

# Find the HuggingFace CLI binary (named 'hf', not 'huggingface-cli')
HF_BIN="$(find "$REPO_DIR/venv" ~/venv/amc -name hf -type f 2>/dev/null | head -1)"
{ [ -z "$HF_BIN" ] || [ ! -x "$HF_BIN" ]; } && { echo "ERROR: hf binary not found or not executable; install the hf CLI (Step 0c) or set HF_BIN" >&2; exit 1; }

# Do NOT use --token on the command line (leaks via ps/argv). The HF CLI
# reads HF_TOKEN from the environment automatically.
HF_TOKEN="<HF_TOKEN>" "$HF_BIN" download facebook/VGGT-1B-Commercial \
  --local-dir models/vggt/

# Verify
ls -lh models/vggt/vggt_1B_commercial.pt
# Should show ~4.7GB file

Important: Download BEFORE setting chown 1000:1000 on the models directory — the current user needs write access during download. Set permissions in Step 4 after download completes.

Step 3: Configure Compose Environment Variables

The Compose environment file controls ports and paths. Update it before launching:

cd $REPO_ROOT/compose

# Find available backend port (8000-8009)
for port in {8000..8009}; do
  if ! lsof -Pi :$port -sTCP:LISTEN -t >/dev/null 2>&1; then
    MS_PORT=$port
    echo "Using backend port: $MS_PORT"
    break
  fi
done
[ -z "$MS_PORT" ] && { echo "ERROR: no free backend port in 8000-8009; free one or widen the range." >&2; exit 1; }

# Find available UI port (5000-5009)
for port in {5000..5009}; do
  if ! lsof -Pi :$port -sTCP:LISTEN -t >/dev/null 2>&1; then
    UI_PORT=$port
    echo "Using UI port: $UI_PORT"
    break
  fi
done
[ -z "$UI_PORT" ] && { echo "ERROR: no free UI port in 5000-5009; free one or widen the range." >&2; exit 1; }

# Get host IP
HOST_IP=$(hostname -I | awk '{print $1}')
echo "Host IP: $HOST_IP"

# Preserve existing keys and restrict permissions on the Compose environment file.
COMPOSE_ENV_BASENAME="env"
ENV_FILE=".${COMPOSE_ENV_BASENAME}"
if [ -f "$ENV_FILE" ]; then
  BACKUP="${ENV_FILE}.bak.$(date +%s)"
  cp "$ENV_FILE" "$BACKUP"
  chmod 600 "$BACKUP"
fi
touch "$ENV_FILE"
chmod 600 "$ENV_FILE"
set_env_key() {
  local k="$1" v="$2"
  if grep -qE "^${k}=" "$ENV_FILE"; then
    sed -i "s|^${k}=.*|${k}=${v}|" "$ENV_FILE"
  else
    echo "${k}=${v}" >> "$ENV_FILE"
  fi
}
set_env_key AUTO_MAGIC_CALIB_MS_PORT "${MS_PORT}"
set_env_key AUTO_MAGIC_CALIB_UI_PORT "${UI_PORT}"
set_env_key PROJECT_DIR "../../projects"
set_env_key MODEL_DIR "../../models"
set_env_key HOST_IP "${HOST_IP}"

# Keep timestamped Compose environment backups out of git.
GITIGNORE="$REPO_ROOT/.gitignore"
touch "$GITIGNORE"
BACKUP_PATTERN="compose/${ENV_FILE}.bak.*"
grep -qxF "$BACKUP_PATTERN" "$GITIGNORE" || echo "$BACKUP_PATTERN" >> "$GITIGNORE"

echo "✓ Compose environment file updated"
cat "$ENV_FILE"

Important: HOST_IP must be the machine's network IP (not localhost) so the UI container can reach the backend from a browser.

Optional: set VGGT_MODEL_PATH only if the VGGT model is mounted at a non-default container path; default is /tmp/vggt_model/vggt_1B_commercial.pt inside the MS container.

Optional for RTSP calibration: use skills/amc-run-rtsp-calibration/SKILL.md after launch. That skill verifies VIOS reachability and, when needed, relaunches the microservice with a temporary compose override that exports VIOS_BASE_URL without changing checked-in compose files.

Step 4: Set Directory Permissions

The containers run as UID/GID 1000. The projects and models directories must be owned by this UID for containers to read/write properly:

cd "$REPO_ROOT"

# Create projects directory if it doesn't exist
mkdir -p projects

# Set ownership (required for containers to write calibration outputs).
# Do this AFTER VGGT download is complete (current user needs write access during download).
# Get explicit user confirmation before running sudo chown — it recursively changes
# ownership of $REPO_ROOT/projects and $REPO_ROOT/models to UID/GID 1000.
[ -d projects ] && [ -d models ] || {
  echo "ERROR: expected projects/ and models/ under $REPO_ROOT" >&2; exit 1;
}
echo "About to chown -R 1000:1000 on:"
echo "  $REPO_ROOT/projects"
echo "  $REPO_ROOT/models"
echo "(required because containers run as UID 1000). Confirm before proceeding."
sudo chown 1000:1000 -R projects
sudo chown 1000:1000 -R models

echo "✓ Permissions set"

Step 5: Launch Services

Before pulling, fail fast if the NGC key authenticated in Step 1 but cannot actually access a release image — otherwise docker compose up aborts partway with a 401/403 after some work is already done.

cd $REPO_ROOT/compose

# Fail-fast image-access check: confirm the NGC key can reach every release
# image BEFORE pulling. `docker manifest inspect` checks registry access without
# downloading layers, and the image list is read from the resolved compose so it
# tracks the release tag automatically.
IMAGES=$(docker compose config --images | sort -u)
[ -z "$IMAGES" ] && { echo "ERROR: no images resolved from compose — check the Compose environment settings and chosen profile." >&2; exit 1; }
for img in $IMAGES; do
  echo "Checking access: $img"
  if ! docker manifest inspect "$img" >/dev/null 2>&1; then
    echo "NGC login succeeded, but this key cannot access the required image:" >&2
    echo "  $img" >&2
    echo "Provide an NGC key with access to this image's namespace, then re-run Step 1 (login) and retry." >&2
    exit 1
  fi
done

# Start all services (images pulled automatically on first run)
docker compose up -d

# Check containers are running
docker compose ps

The exact image tags change by release; read them from the active compose files instead of hardcoding a version.

Step 6: Verify Services Are Running

# Read ports from the Compose environment file.
COMPOSE_ENV_BASENAME="env"
COMPOSE_ENV_FILE="$REPO_ROOT/compose/.${COMPOSE_ENV_BASENAME}"
MS_PORT=$(grep AUTO_MAGIC_CALIB_MS_PORT "$COMPOSE_ENV_FILE" | cut -d= -f2)
UI_PORT=$(grep AUTO_MAGIC_CALIB_UI_PORT "$COMPOSE_ENV_FILE" | cut -d= -f2)
HOST_IP=$(grep HOST_IP "$COMPOSE_ENV_FILE" | cut -d= -f2)

# Wait for microservice readiness. Cold image pulls or first startup can need
# extra time after `docker compose up -d` returns.
READY_URL="http://localhost:${MS_PORT}/v1/ready"
echo "Waiting for microservice readiness at ${READY_URL} ..."
ready_response=""
for attempt in $(seq 1 24); do
  if ready_response=$(curl -fsS --max-time 5 "${READY_URL}" 2>/dev/null) && \
     echo "${ready_response}" | grep -q '"code"[[:space:]]*:[[:space:]]*0'; then
    echo "Microservice ready: ${ready_response}"
    break
  fi
  if [ "${attempt}" -lt 24 ]; then
    printf "  [%02d/24] Microservice not ready yet; retrying in 5s...\n" "${attempt}"
    sleep 5
  fi
done

if ! echo "${ready_response}" | grep -q '"code"[[:space:]]*:[[:space:]]*0'; then
  echo "ERROR: microservice did not report ready within 120 seconds: ${READY_URL}" >&2
  echo "Check status and logs:" >&2
  echo "  cd ${REPO_ROOT}/compose && docker compose ps" >&2
  echo "  cd ${REPO_ROOT}/compose && docker compose logs auto-magic-calib-ms" >&2
  exit 1
fi

# Check UI is serving
UI_STATUS=$(curl -s -o /dev/null -w "%{http_code}" --max-time 5 "http://localhost:${UI_PORT}")
if [ "${UI_STATUS}" != "200" ]; then
  echo "ERROR: Web UI returned HTTP ${UI_STATUS}; check docker compose ps and UI logs." >&2
  exit 1
fi
echo "Web UI ready: HTTP ${UI_STATUS}"

echo "Microservice: http://${HOST_IP}:${MS_PORT}"
echo "Web UI:       http://${HOST_IP}:${UI_PORT}"

Success Criteria

  • docker compose ps shows MS and UI containers Up; MS should be healthy.
  • /v1/ready returns code:0 and Step 6 prints the microservice and UI URLs.
  • Browser access to http://<HOST_IP>:<AUTO_MAGIC_CALIB_UI_PORT> works.
  • Projects persist under $REPO_ROOT/projects/.

Troubleshooting

IssueFix
Docker permission deniedAsk the user to run sudo usermod -aG docker $USER && newgrp docker, then retry docker ps.
docker login rejectedAsk for a current NGC key and log in again.
Required image inaccessibleThe key lacks image namespace access; ask for a key with access, then retry Step 1 and Step 5.
python3 -m venv, pip, or hf missingInstall python3-venv/python3-pip; the HF binary is named hf.
VGGT permission errorDownload VGGT before chown 1000:1000; to recover, restore user ownership of models/ and re-download.
Port in usePick a free MS port in 8000-8009 and UI port in 5000-5009, then update the Compose environment file.
Readiness timeout or exited containerRun cd $REPO_ROOT/compose && docker compose ps and inspect docker compose logs auto-magic-calib-ms.
Project/model permission deniedRe-run Step 4 for projects/ or models/ only.
UI cannot reach backendVerify HOST_IP in the Compose environment file is the machine network IP, not localhost.
GPU unavailableVerify NVIDIA runtime with docker run --rm --runtime=nvidia --gpus all ubuntu:20.04 nvidia-smi.

Common Fixes:

cd $REPO_ROOT/compose

# View logs
docker compose logs -f

# View logs for specific service
docker compose logs -f auto-magic-calib-ms

# Restart all services
docker compose restart

# Stop and remove containers
docker compose down

# Update Compose environment settings and relaunch
docker compose up -d

Stopping the Services

cd $REPO_ROOT/compose

# Stop all services (containers removed, data persisted)
docker compose down

# Stop and remove volumes
docker compose down -v

Related Skills

  • skills/amc-run-sample-calibration/SKILL.md - Sanity-check the running stack with the bundled sample dataset
  • skills/amc-run-video-calibration/SKILL.md - Calibrate from your own pre-recorded MP4s via REST API
  • skills/amc-run-rtsp-calibration/SKILL.md - Calibrate from live RTSP streams through VIOS capture

Plus de skills de nvidia

compileiq-debug
nvidia
Utilisez quand quelque chose ne va pas : Search() bloque, toutes les évaluations retournent INVALID_SCORE, les scores ne s'améliorent pas, chaque configuration retourne le même nombre, erreurs ptxas…
create-github-pr
nvidia
Créer des pull requests GitHub en utilisant l'interface en ligne de commande gh. Utiliser lorsque l'utilisateur souhaite créer une nouvelle PR, soumettre du code pour révision, ou ouvrir une pull request. Mots-clés de déclenchement -…
nemoclaw-maintainer-cross-issue-sweep
nvidia
Analyse les autres problèmes ouverts pour trouver ceux qu’une PR donnée pourrait également corriger ou casser accidentellement. Génère des opportunités de correctifs adjacents et des risques de contradiction avec fichier:ligne…
fhir-basics
nvidia
Apprend aux agents comment fonctionnent les API FHIR R4, quelles ressources sont disponibles, comment les interroger avec des paramètres de recherche, et comment analyser correctement tous les formats de réponse…
compileiq-validate-result
nvidia
Utiliser APRÈS qu'une recherche soit terminée et AVANT de réclamer un accélérateur ou d'expédier un ACF. Charge le CSV dump_results, extrait les K meilleurs candidats (mono-objectif)…
changelog-audit
nvidia
Auditer le CHANGELOG.md de Warp avant une publication : récupérer les entrées perdues, trier par impact utilisateur, affiner le langage des entrées, ajuster les retours à la ligne et (en mode branche de publication) mettre à jour la comparaison…
maintain-dynamic-plugins
nvidia
Maintenir les chargeurs de plugins dynamiques NeMo Relay, les manifestes, les SDK natifs Rust, le protocole worker gRPC, le SDK worker Python, la documentation, les tests et la couverture du workflow de publication
dgx-diagnose
nvidia
Diagnostiquer les problèmes courants du DGX Station GB300 — plantages CUDA, ciblage incorrect du GPU, bugs de conteneur vLLM/SGLang, problèmes d'état MIG, erreurs NVLink/Fabric Manager,…