+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Evaluation & Testing (https://www.datasculpting.ai/inference/forumdisplay.php?fid=16)
+--- Forum: Model Comparisons (https://www.datasculpting.ai/inference/forumdisplay.php?fid=67)
+--- Thread: image generation model locals actually use (/showthread.php?tid=360)
Pages:
1
2
Re: image generation model locals actually use - sparrowshed - 09-28-2026
I'd approach it in a boring order. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at token cost.
Re: image generation model locals actually use - phasescope - 09-28-2026
Thanks, that helps. The pattern around prefix caching has been pretty similar for me.
Re: image generation model locals actually use - mlxgarden - 09-29-2026
Appreciate the detail. I ended up in about the same place using llama.cpp.
Re: image generation model locals actually use - kernelghost - 09-29-2026
What worked for me was pretty simple. If you're on MLX, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.