IQ2_XXS
PulseAugur coverage of IQ2_XXS — every cluster mentioning IQ2_XXS across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
llama.cpp adds OpenCL support for IQ2 quantization on Adreno GPUs
The llama.cpp project has released version b11986, introducing initial OpenCL support for IQ2 quantization formats. This update includes optimizations and specific implementations for Adreno GPUs, targeting IQ2_S, IQ2_X…
-
How to Fairly Compare Quantized LLMs in llama.cpp
This article provides a guide on how to accurately compare quantized Large Language Models (LLMs) using llama.cpp. It emphasizes the importance of matching file sizes in bytes, not just quantization labels, as labels ca…
-
Unsloth trims Kimi K3 LLM to under 500GB, boosting performance
Unsloth has released several versions of the Kimi K3 large language model, significantly reducing their file sizes. One version, IQ2_XXS, was trimmed from 711GB to 478GB by removing multilingual capabilities and retaini…
-
User releases 140GB GLM-5.2 coding quantization for testing
A user has created a 140 GB IQ2_XXS REAP quantization of the GLM-5.2 model, specifically optimized for coding tasks. This quantization is based on the "most liked REAP" and is available on Hugging Face. The creator is s…