seeed-projects/seeed-jetson-developtool/generative-ai-intro

Generative AI with reComputer Jetson

Execution model

Run one phase at a time. After each phase:

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

Prerequisites

Phase 1 — Identify the target application

This skill is an index of Generative AI topics on Jetson. Help the user identify which specific application they need:

Text Generation (Local Chatbots)

• Local AI Assistant with Ollama + AnythingLLM
• Local Voice Chatbot with Nvidia Riva + Llama2

Image Generation

• Local LLM Text-to-Image on reComputer

Audio Generation

• Speech Subtitle Generation on Jetson
• Deploy Whisper on Jetson Orin for real-time STT

Multimodal

• Run VLM on reComputer with Jetson Platform Services

Retrieval Augmented Generation (RAG)

• Local AI Assistant with Ollama + AnythingLLM
• Local RAG with LlamaIndex on Jetson

Other

• Fine-tune LLM with Llama-Factory on Jetson
• Quantized Llama2-7B with MLC LLM on Jetson Orin NX
• Zero-Shot Detection on reComputer
• Format LLM Output with Langchain on Jetson

[OK] — once the user selects a topic, direct them to the corresponding skill or wiki link.

Phase 2 — Deploy jetson-examples (common entry point)

Many Generative AI applications on Jetson use the jetson-examples package for one-line deployment:

pip3 install jetson-examples
sudo reboot

After reboot, deploy the chosen application:

reComputer run <application-name>

[OK] when the application container starts and the service is accessible.

For the full list of applications, wiki links, and detailed tutorials for each topic, see references/source.body.md.

Failure decision tree

Reference files

• references/source.body.md — full original Seeed index page with links to all Generative AI tutorials, screenshots, and application descriptions (reference only)