huihui-ai/Qwen2.5-3B-Instruct-abliterated-SFT
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:3.1BQuant:BF16Ctx Length:32kPublished:Apr 13, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

huihui-ai/Qwen2.5-3B-Instruct-abliterated-SFT is an instruction-tuned language model developed by huihui-ai, fine-tuned from the Qwen2.5-3B-Instruct-abliterated base model. This model is specifically trained on the Guilherme34_uncensor dataset, focusing on uncensored content generation. It is designed for conversational AI applications requiring less restrictive content filtering, leveraging the Qwen2.5 architecture.

Loading preview...

Model Overview

huihui-ai/Qwen2.5-3B-Instruct-abliterated-SFT is an instruction-tuned language model developed by huihui-ai. It is a specialized variant of the Qwen2.5-3B-Instruct-abliterated model, further fine-tuned to address specific content generation requirements.

Key Characteristics

  • Base Model: Fine-tuned from huihui-ai/Qwen2.5-3B-Instruct-abliterated.
  • Training Data: Utilizes the huihui-ai/Guilherme34_uncensor dataset for its instruction tuning, indicating a focus on generating less restricted or uncensored content.
  • License: Distributed under the Apache-2.0 license.
  • Implementation: The README provides detailed Python code snippets for loading and interacting with the model using the transformers library, including configurations for BitsAndBytesConfig and a custom TextStreamer for interactive chat.

Intended Use Cases

This model is particularly suited for applications where the goal is to generate responses with fewer content restrictions, as implied by its training on an "uncensor" dataset. Developers can integrate it into conversational agents or content generation systems that require a broader range of expression than typically allowed by more heavily moderated models. The provided code demonstrates a chat interface, suggesting its utility in interactive dialogue systems.