All issues

Search results

gguf

Status: Open.

#12 In IST-DASLab/GSQ;

· ayttop opened on Sep 9, 2026

-mtp GGUFs: IQ3_S-mtp MTP head gets 0.0% draft acceptance (XXS-mtp works)

Status: Open.

#11 In IST-DASLab/GSQ;

· sunyan999999 opened on Sep 9, 2026

OpenThoughts preprocessing hardcodes a Kimi-specific system prompt

Status: Open.

#10 In IST-DASLab/GSQ;

· anzr299 opened on Sep 9, 2026

Reproduction tooling for Qwen3.8-27B GSQ-RCO mixed-precision GGUF release

Status: Open.

#9 In IST-DASLab/GSQ;

· SoppyLeon opened on Sep 7, 2026

SM80 (Ampere/A100/170HX) GSQ INT3 decode speed bottlenecked by upstream Humming kernel, not GSQ quality

Status: Open.

#8 In IST-DASLab/GSQ;

· yangfangyunyun opened on Sep 6, 2026

NPROC=2 OOMs on 2x12GB consumer GPUs where NPROC=1 succeeds (Qwen3.6-27B dense, gsq_bits=2); expandable_segments:True also masks OOM as "device not ready" on WSL2

Status: Open.

#6 In IST-DASLab/GSQ;

· Nakanokensetsu opened on Jul 3, 2026

How to refine a public GGUF K-Quant model?

Status: Open.

#5 In IST-DASLab/GSQ;

· yancaoweidaode opened on Jul 1, 2026