zypchn/BehChat-qwen-SFT-v2

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:7.6BQuant:FP8Ctx Length:32kPublished:May 29, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

BehChat-qwen-SFT-v2 is a 7.6 billion parameter Qwen2-based causal language model developed by zypchn, fine-tuned from unsloth/deepseek-r1-distill-qwen-7b-unsloth-bnb-4bit. This model leverages Unsloth for accelerated training, offering a 32768 token context length. It is optimized for chat and instruction-following tasks, benefiting from efficient fine-tuning techniques.

Loading preview...

BehChat-qwen-SFT-v2 Overview

BehChat-qwen-SFT-v2 is a 7.6 billion parameter instruction-tuned language model developed by zypchn. It is based on the Qwen2 architecture and was fine-tuned from the unsloth/deepseek-r1-distill-qwen-7b-unsloth-bnb-4bit model, utilizing Unsloth for a 2x speedup in training and Huggingface's TRL library.

Key Capabilities

  • Efficiently Trained: Leverages Unsloth for faster fine-tuning, making it a resource-efficient option for deployment.
  • Qwen2 Architecture: Benefits from the robust capabilities of the Qwen2 model family.
  • Instruction Following: Designed for chat and instruction-based interactions due to its supervised fine-tuning (SFT).
  • Extended Context: Supports a context length of 32768 tokens, allowing for processing longer inputs and maintaining conversational coherence over extended dialogues.

Good For

  • Chatbots and Conversational AI: Its instruction-tuned nature makes it suitable for building interactive agents.
  • General Text Generation: Capable of generating coherent and contextually relevant text based on prompts.
  • Applications requiring efficient deployment: The model's efficient training suggests it might be optimized for performance, making it a good candidate for applications where speed and resource usage are critical.