Welcome to Jetson Orin NANO Developer Kit repository
08.Google - Gemma visual multimodal large model
Gemma3
Gemma31. Model Size2. Performance3. Using Gemma33.1 Running Gemma33.2 Starting a Conversation3.3 Visual Function3.4 Ending the ConversationReferences
Demo Environment
Development board: Jetson Orin series mainboard
Note: Due to performance limitations, the Jetson Orin Nano 4GB needs to run the small-parameter version
Gemma is a series of lightweight models built by Google based on Gemini technology. The Gemma 3 model is a multimodal model (it can process text and images), has a 128K context window, and supports more than 140 languages.
1. Model Size
| Model | Size |
|---|---|
| gemma3:1b | 815MB |
| gemma3:4b | 3.3GB |
2. Performance
3. Using Gemma3
3.1 Running Gemma3
Run the model with the run command; if it is not downloaded yet, Ollama pulls it automatically:
ollama run gemma3:4b

3.2 Starting a Conversation
xxxxxxxxxxHow to learn a programming language?
Response time depends on the hardware; please be patient.

3.3 Visual Function

xxxxxxxxxxWhat do you see in this picture? :./test_pic.png#Use ": + image path" in the conversation to enable the model to use its visual function and interpret the information in the image

3.4 Ending the Conversation
Use the Ctrl+d shortcut or /bye to end the conversation!
References
Ollama
Official website: https://ollama.com/
GitHub: https://github.com/ollama/ollama
Gemma3
Ollama corresponding model: https://ollama.com/library/gemma3