Hi Bartowski, thank you for all your GGUF work.
I know I’m just a small hobbyist, but I wanted to ask if you might consider looking at Zyphra/ZAYA1-8B. It’s a small MoE model reportedly trained on AMD hardware, and I’m very interested in testing it on edge devices.
I’m building local AI devices for my family and currently run your Gemma-4-E2B GGUFs on a Jetson Orin Nano Super Kit and a Raspberry Pi 5 16GB. The Jetson is fast enough, but the 8GB RAM limit makes small MoE models very interesting for me.
I saw a couple of existing ZAYA1 GGUF quantizations, but they don’t seem to work well yet. I wanted to try my luck and ask if this model is something you might be willing to look at someday.
Thank you very much for everything you do.