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

SpiridonSunRotator authored a paper 11 days ago

FLAMIA: FLAsh MIcroscaling Attention for Fast and Causally Safe Training on Blackwell

SpiridonSunRotator submitted a paper 11 days ago

FLAMIA: FLAsh MIcroscaling Attention for Fast and Causally Safe Training on Blackwell

kwanhee new activity 11 days ago

ISTA-DASLab/Kimi-K3-P48NVFP4-MoESQ:Kimi K3 GSQ-RCO-GGUF IQ1_S, IQ1_M, IQ2_XXS

View all activity

Papers

Hardware-Native Joint Sparse-Quantization for Trillion-Scale Mixture-of-Experts

WUSH-KV: KV Cache Quantization with Data-Adaptive Transforms

View all Papers

SpiridonSunRotator

authored a paper 11 days ago

FLAMIA: FLAsh MIcroscaling Attention for Fast and Causally Safe Training on Blackwell

Paper • 2611.03344 • Published 11 days ago • 21

SpiridonSunRotator

submitted a paper to Daily Papers 11 days ago

FLAMIA: FLAsh MIcroscaling Attention for Fast and Causally Safe Training on Blackwell

Paper • 2611.03344 • Published 11 days ago • 21

in ISTA-DASLab/Kimi-K3-P48NVFP4-MoESQ 11 days ago

Kimi K3 GSQ-RCO-GGUF IQ1_S, IQ1_M, IQ2_XXS

#1 opened about 1 month ago by

12 indexed weight shards missing / 404 EntryNotFound (revision 57ee129)

#2 opened 15 days ago by

updated a model 11 days ago

ISTA-DASLab/Kimi-K3-P48NVFP4-MoESQ

Text Generation • 1.8T • Updated 11 days ago • 489 • 7

in ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF 11 days ago

Unsloth taking shots at ISTA

#29 opened 14 days ago by

updated a collection 11 days ago

GSQ-RCO-GGUF

Non-uniform GGUF quantizations via GSQ + RCO: per-tensor mixed precision in standard GGUF form • 5 items • Updated 11 days ago • 108

published a model 11 days ago

ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-Coder-GGUF

Image-Text-to-Text • 197B • Updated 11 days ago • 403k • 228

updated a model 11 days ago

ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-Coder-GGUF

Image-Text-to-Text • 197B • Updated 11 days ago • 403k • 228

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF 12 days ago

Bad Rust coding quality on new project? (strata)

#28 opened 12 days ago by

in ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF 15 days ago

[Request] Harder benchmarks!

#21 opened 16 days ago by

updated a model 16 days ago

ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF

Text Generation • 197B • Updated 16 days ago • 3.76M • 663

in ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF 17 days ago

Updated chat template

#15 opened 19 days ago by

in ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-Coder-GGUF 17 days ago

I think might have had one too many experts stripped out

#11 opened about 2 months ago by

in ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF 19 days ago

Please align with the unsloth quants

#11 opened 20 days ago by

updated a model 19 days ago

ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF

Text Generation • 197B • Updated 16 days ago • 3.76M • 663

in ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF 19 days ago

Vision support

#8 opened 20 days ago by

IQ2_XS (2.25) worse than IQ2_XXS (2.00) ?

#13 opened 19 days ago by

DzmitryTheOtherOne

updated a collection 20 days ago

GSQ-RCO-GGUF

Non-uniform GGUF quantizations via GSQ + RCO: per-tensor mixed precision in standard GGUF form • 5 items • Updated 11 days ago • 108

in ISTA-DASLab/DeepSeek-V4.1-Flash-1017-GSQ-RCO-GGUF 20 days ago

IQ2_XS vs IQ2XXS - why recommend the smaller one?

#3 opened 21 days ago by