criscarleo/qwen2.5-coder-3b-abliterated-basic

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:3.1BQuant:BF16Ctx Length:32kPublished:Mar 23, 2026License:agpl-3.0Architecture:Transformer Open Weights Warm

The criscarleo/qwen2.5-coder-3b-abliterated-basic model is a Qwen2.5-based language model processed using the Obliteratus methodology. This transformation aims to optimize the model for specific tasks, leveraging the original Qwen2.5 architecture. It is designed for efficient deployment and inference, particularly with the GGUF format via llama.cpp. The model's primary utility lies in applications benefiting from its specialized processing and compact format.

Loading preview...

Model Overview

The criscarleo/qwen2.5-coder-3b-abliterated-basic model is a variant of the Qwen2.5 architecture that has undergone a specialized transformation using the Obliteratus methodology. This process, developed by pliny-the-prompter, aims to refine and optimize the base model for enhanced performance or specific characteristics.

Key Characteristics

  • Transformation Method: Utilizes the Obliteratus methodology, a specialized fine-tuning/transformation process.
  • Tooling: The transformation was executed via a Jupyter Notebook implementation provided by the Obliteratus Space.
  • Quantization & Inference: Optimized for efficient deployment, packaged in the GGUF (v3) format.
  • Framework: Designed for use with the llama.cpp inference engine, known for its performance on consumer hardware.

Intended Use Cases

This model is particularly suitable for:

  • Applications requiring a Qwen2.5-based model with specific optimizations from the Obliteratus process.
  • Edge device deployment or scenarios where efficient inference with llama.cpp and GGUF format is crucial.
  • Developers interested in exploring models processed with the Obliteratus methodology for potential performance or capability enhancements.