#728 Infinite loop issue

open · @davidliudev · 18 commentaires · Sur GitHub

Benchmarks Setup & install Models & quants Windows

Description

Having switch-1.50iq2_xs, 8 bit KV cache, 128k context size, engine version v0.1.38 running out of the box after install on 4090 and 64 GB RAM, Windows 11.

Tried OpenCode and Qwen Coder but both will enter endless loop after a while.
Performance is good at 103 tok/s except when this issue was hit.

Is there anything I can tweak to prevent this?

Sur le site

Install Strata
Model sizes & RAM
Release notes

Liens install, modèles, releases.