https://stephenkrings.com/tags/hugging-face/ Recent content in Hugging-Face on Stephen Krings Hugo en-us Tue, 14 Jul 2026 19:58:00 -0600 https://stephenkrings.com/posts/hugging-face-image-generation-rtx-5070-cpu-offload/ Tue, 14 Jul 2026 19:58:00 -0600 https://stephenkrings.com/posts/hugging-face-image-generation-rtx-5070-cpu-offload/ <p>Goal: wiring Hugging Face models into my local llm (Ollama) app, testing prompt-following, and seeing how far I can push a 12GB RTX 5070 before the system starts leaning hard on the CPU.</p> <p>The short version: it works, but only because the stack is allowed to spill parts of the model back to system memory.</p> <h2 id="the-current-setup">The current setup</h2> <p>The image UI lives on my web server, but the GPU does not. The request path looks like this:</p>