Machine LearningDev30 min reading time

How to Run Underdog Saluki 27B: 2-Bit Qwen3.8-27B Tool-Calling Guide

MarkTechPost
Read the full article
Underdog Saluki 27B is a 7.9 GB quantized version of Qwen3.8-27B that maintains tool-calling capabilities at 2-bit precision and outperforms the full 54 GB model on a 120-task benchmark. The tutorial compares Saluki to ISTA-DASLab's GSQ-RCO quantization using llama.cpp, evaluating memory, speed, and task performance.

More in Machine Learning

Machine Learning5 min read

PyTorch Conference 2026: Google and Meta Open-Source TorchTPU for Native PyTorch on TPUs

InfoQ (AI, ML & Data)
Machine Learning5 min read

Hugging Face Releases Carbon-X, a 14B Open DNA Language Model that Brings DNA and Natural Language Together for the First Time

MarkTechPost
Machine Learning3 min read

Google ML Drift Accelerates On-Device AI Inference Across Diverse GPUs

InfoQ (AI, ML & Data)