AMD Rolls Out Gemma 4 Model Support Across Full Range of GPUs & CPUs
wccftech.com • • technology
Key Points:
- AMD has announced full support for Google's Gemma 4 AI model family across its entire Radeon GPU and Ryzen AI CPU lineup, including AMD Instinct GPUs for datacenters and Ryzen AI processors for PCs.
- The Gemma 4 models, ranging from 2 billion to 31 billion parameters, can be deployed on AMD hardware using popular AI frameworks such as vLLM and SGLang, with optimizations for concurrent requests and high-performance serving.
- AMD enables easy local deployment of Gemma 4 models on Ryzen AI and Radeon GPUs through open-source tools like llama.cpp, LM Studio, and Lemonade Server, supporting acceleration via ROCm and the XDNA 2 NPU.
- Lemonade Server provides OpenAI-compatible APIs for running Gemma 4 on AMD GPUs and NPUs, with upcoming updates to support NPU deployment on Ryzen AI processors through OnnxRuntime APIs and integrated software updates.
- Support includes integration with widely used AI applications and open-source projects, ensuring broad accessibility and optimized performance on AMD’s AI-enabled hardware portfolio.
