This Hugging Face repository provides a W4A16 AWQ/GPTQ quantization of a Qwen 27B model for lower-memory inference. Developers can download and run the weights with compatible local inference tools or deploy them in self-hosted environments.
In the Quantised & converted weights space, Qwen3.8 27B W4A16 takes a focused approach. It focuses on running a large Qwen model with reduced memory requirements for local or self-hosted inference. Qwen3.8 27B W4A16 is an open-source project aimed at machine learning engineers and developers. The project is open source (Apache-2.0). It ships for the web, the command line, and API, and it can be self-hosted.
Behind Qwen3.8 27B W4A16 is soyrsoyr, and it first shipped in 2019. Development happens publicly on GitHub with 3.8k stars and 179 commits in the last 90 days. The category is crowded — PulseGate's index counts 25 comparable projects. Key capabilities include quantized weights, AWQ format, and GPTQ format.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do