Smaller quantized files make a reasoning model easier to run locally, but compression choices can affect output quality. This repository offers multiple imatrix-weighted GGUF variants of a decensored 9B model, letting users choose a practical size-quality trade-off rather than relying on a single quantization.
AI Model2026
mradermacher/DeepSeek-V4-Pro-Qwen3.5-9B-UNCENSORED-heretic-i1-GGUF
Provides imatrix-weighted GGUF quantizations of a 9B model for local inference, with options balancing file size and quality. The source is a Heretic-decensored model focused on mathematical and STEM reasoning.
Introduction
Information
- Websitehuggingface.co
- Authorsmradermacher
- Published date2026/10/31
Categories
More Items
AI Model2026
Runs GLM-5.3 as a 192-expert, roughly 2.56-bit GGUF model on a single NVIDIA B200. It is 24.6% smaller than the 256-expert quantized version, trading some model capacity for a smaller local-inference footprint.
AI Image2026
Turns text prompts or up to eight reference images into high-resolution images, with support for editing. An eight-step sampler and built-in prompt upsampler streamline generation, though bf16 inference requires substantial GPU memory.
