AMD’s Radeon RX 7900 XTX clearly stands out when running the DeepSeek R1 AI model, even surpassing NVIDIA’s GeForce RTX 4090 in inference benchmarks.
## AMD’s Quick Response Enhances DeepSeek R1 Model Performance
DeepSeek’s latest AI model is making waves in the industry, and while many are curious about the computational power required for its training, it turns out that ordinary users can achieve impressive results with the Radeon RX 7900 XTX GPU from AMD’s “RDNA 3” lineup. AMD has released comparison benchmarks that pit their flagship RX 7000 series GPU against NVIDIA’s offering, showing clear advantages across different models.
> DeepSeek performing excellently on @AMDRadeon 7900 XTX. Discover how to run on Radeon GPUs and Ryzen AI APUs here: https://wccftech.com/amd-radeon-rx-7900-xtx-beats-nvidia-geforce-rtx-4090-in-deepseeks-ai-inference-benchmark/ — David McAfee (@McAfeeDavid_AMD) January 29, 2025
For many users, consumer GPUs have been a preferred choice for AI tasks, largely due to their favorable performance-to-cost ratio when compared to typical AI accelerators. Running AI models locally also means enhanced privacy, a significant concern with DeepSeek’s AI models. Luckily, AMD has developed a comprehensive guide on executing DeepSeek R1 distillations using their GPUs. Here’s a quick rundown of the steps:
1. Ensure you have the 25.1.1 Optional or higher Adrenalin driver.
2. Download LM Studio 0.3.8 or later from lmstudio.ai/ryzenai.
3. Install LM Studio and bypass the onboarding screen.
4. Go to the discover tab.
5. Select your DeepSeek R1 Distill. Start with smaller options like the fast Qwen 1.5B, although larger models provide enhanced reasoning abilities.
6. Make sure “Q4 K M” quantization is selected on the right-hand side, then click “Download.”
7. After downloading, go to the chat tab, pick the DeepSeek R1 distill from the menu, and ensure “manually select parameters” is checked.
8. Slide the GPU offload layers to the maximum setting.
9. Click on model load.
10. Now, interact with the reasoning model running completely using your local AMD hardware!
If you’re having trouble with these instructions, AMD has uploaded a detailed YouTube tutorial that breaks down each step. It’s worth having a look to ensure DeepSeek’s language models are operating securely on your own AMD equipment. As new NVIDIA and AMD GPUs hit the market, expect a huge leap in inference capabilities, thanks to integrated AI engines designed for these tasks.
