gemma3

Gemma3 (4B)

This directory contains the docker configuration files to run Gemma3-4B on RyzenAI platforms. Gemma3-4B is the smallest open-weights model released under the Gemma3 suite that offers multimodal support.

Build and run the Docker Image

To build and run a Docker container with Gemma3-4B-Instruct using a llamacpp backend, run:

ryzers build llamacpp gemma3
ryzers run

Demo

Additionally, there is a demo.sh included that when run on the host machine, will serve up a webpage that allows a user to interact with Gemma3-4B VLM. A webcam is required. Credit to https://github.com/ngxson/smolvlm-realtime-webcam.

git clone https://github.com/ngxson/smolvlm-realtime-webcam <PATH TO REPO>

ryzers build llamacpp gemma3
ryzers run /ryzers/demo_gemma3.sh

In a browser, open file://<PATH TO REPO>/index.html

Build and run with HuggingFace Transformers

gemma3 can also be run with the transformers API. To do so, run:

ryzers build gemma3
ryzers run bash

# inside docker container
cd /ryzers
HSA_OVERRIDE_GFX_VERSION=11.0.0 HF_TOKEN=<your_huggingface_token> python test_gemma3_hf.py

Copyright(C) 2025 Advanced Micro Devices, Inc. All rights reserved.