Peter Dedene
@dedene
A dense 27B model in 7.89 GB, running at ~22 tok/s on my DGX Spark with 262k context. And 87 GB of memory still free. 🤯
Saluki is pretty wild.
If you want to try it yourself:
github.com/dedene/Underdo…
→ docker compose up
Saluki is pretty wild.
If you want to try it yourself:
github.com/dedene/Underdo…
→ docker compose up
github.comGitHub - dedene/Underdog-Saluki-27B-DGX-Spark
Underdog AI@UnderdogAI · Oct 8Today we're releasing Underdog Saluki 27B
Based on Qwen 3.8 27B, Saluki is almost 7x smaller than its full-precision counterpart while it keeps 96% of its benchmark performance and beats Qwen at tool calling
It's the best model <8GB that runs in Underdog
Saluki is available
Based on Qwen 3.8 27B, Saluki is almost 7x smaller than its full-precision counterpart while it keeps 96% of its benchmark performance and beats Qwen at tool calling
It's the best model <8GB that runs in Underdog
Saluki is available
Open quoted post →
3 27