deciai-setup

Getting Started with Deci AI on Jetson

Deci accelerates deep-learning inference 2–10× on any hardware via its AutoNAC compiler and INFERY runtime. This skill walks through account setup, model zoo exploration, custom model optimization, and on-device benchmarking on Jetson.

---

Execution model

Run one phase at a time. After each phase:

• Relay all output to the user.
• If output contains [STOP] → stop immediately, consult the failure decision tree.
• If output ends with [OK] → tell the user "Phase N complete" and proceed to the next phase.

---

Prerequisites

Requirement	Detail
Jetson device	Nano, Xavier NX, TX2, AGX Xavier, or AGX Orin (Seeed reComputer or NVIDIA devkit)
JetPack	4.6
Python	3.6+ with pip3
Network	Internet access for Deci platform and NGC
Deci account	Free sign-up at https://console.deci.ai/sign-up

---

Phase 1 — System update and pip setup (~2 min)

sudo apt update
sudo apt install -y python3-pip
python3 -m pip install -U pip

[OK] when pip3 reports its version. [STOP] if apt or pip fails.

---

Phase 2 — Create Deci account (manual, ~3 min)

1. Visit https://console.deci.ai/sign-up
2. Complete the registration form
3. Log in to Deci Lab — you should see the dashboard with the default ResNet50 Baseline model

[OK] when the user confirms they can see the Deci Lab dashboard.

---

Phase 3 — Explore Model Zoo and optimize a model (manual, ~5–10 min)

In Deci Lab:

1. Click Model Zoo → List to browse pre-optimized models
2. Search for a model (e.g. YOLOX) to see base and hardware-optimized variants
3. To optimize a custom model: click + New Model → select task → upload ONNX model → choose target hardware → set throughput goal → click Start
4. Wait for optimization to complete

[OK] when the optimized model appears in the dashboard.

---

Phase 4 — Install INFERY on Jetson (~3 min)

sudo python3 -m pip install https://deci-packages-public.s3.amazonaws.com/infery_jetson-3.2.2-cp36-cp36m-linux_aarch64.whl