IST Austria Distributed Algorithms and Systems Lab

university

https://ist.ac.at/en/research/alistarh-group/

https://github.com/IST-DASLab

AI & ML interests

None defined yet.

Recent Activity

anm2211 new activity about 16 hours ago

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF:Personal test summary — it falls short of the official version in performance, but is excellent as an IQ3 quantization.

anm2211 new activity about 19 hours ago

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF:IQ3_XXS transformer + higher-precision Q8/BF16 n-gram in system RAM?

anm2211 new activity about 19 hours ago

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF:why no mtp heads

View all activity

Papers

GPTQ-2D: Cubic-Time Two-Sided Adaptive Rounding

MatryoshkaLoRA: Learning Accurate Hierarchical Low-Rank Representations for LLM Fine-Tuning

View all Papers

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF about 16 hours ago

Personal test summary — it falls short of the official version in performance, but is excellent as an IQ3 quantization.

#14 opened 1 day ago by

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF about 19 hours ago

IQ3_XXS transformer + higher-precision Q8/BF16 n-gram in system RAM?

#15 opened 1 day ago by

why no mtp heads

#16 opened about 23 hours ago by

Can I replace per_layer_token_embd.weight in GGUF with its original BF16 weight?

#17 opened about 21 hours ago by

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 1 day ago

IQ3_XXS stops after ~70 tokens on llama.cpp & RTX 5060ti CUDA toolkit 13.2.78

#10 opened 2 days ago by

in ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF 2 days ago

Swift-Qwen3.8-27B-GGUF - is it possible to quantize this model?

#37 opened 4 days ago by

Anyone unable to load the model on LLama.cpp?

#38 opened 3 days ago by

Iloveprogramming342

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 2 days ago

Any plans for larger Flash-Next builds (up to 4 bpw?)

#6 opened 2 days ago by

Amazing. Can only wish for your quants of GLM 5.3 flash

#7 opened 2 days ago by

what should be the parametric setting to get the model tunned for STEM [GPQA-Diamond/LCB/AIME]

#8 opened 2 days ago by

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF 3 days ago

MTP support?

#3 opened 3 days ago by

Q1 version

#5 opened 3 days ago by

in ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF 3 days ago

request for qwen3.8 next

#3 opened 19 days ago by

updated a collection 3 days ago

GSQ-RCO-GGUF

Non-uniform GGUF quantizations via GSQ + RCO: per-tensor mixed precision in standard GGUF form • 2 items • Updated 3 days ago • 27

published a model 3 days ago

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF

Image-Text-to-Text • 177B • Updated 3 days ago • 21.2k • 127

updated a model 3 days ago

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF

Image-Text-to-Text • 177B • Updated 3 days ago • 21.2k • 127

in ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF 5 days ago

Template: resolved_reasoning_effort for `medium` is missing? XXS

#34 opened 5 days ago by

bluegorilla4192

Model outputs garbage in Hermes Agent

#35 opened 5 days ago by

in ISTA-DASLab/Qwen3.8-27B-GSQ-RCO-GGUF 7 days ago

IQ3_S-mtp failed on simple Math

#28 opened 7 days ago by

how to use imatrix-qwen3.8-27b.gguf in llama.cpp

#27 opened 7 days ago by