kmd2525/dpo-qwen-cot-merged
Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Feb 13, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

The kmd2525/dpo-qwen-cot-merged model is a 4 billion parameter language model, fine-tuned from Qwen/Qwen3-4B-Instruct-2507 using Direct Preference Optimization (DPO). It is specifically optimized to enhance reasoning capabilities through Chain-of-Thought (CoT) and improve the quality of structured responses. This model is suitable for applications requiring improved logical coherence and well-formed outputs.

Loading preview...

Model Overview

This model, kmd2525/dpo-qwen-cot-merged, is a 4 billion parameter language model derived from the Qwen/Qwen3-4B-Instruct-2507 base model. It has undergone fine-tuning using Direct Preference Optimization (DPO) via the Unsloth library, with its full 16-bit weights merged for direct use without adapters.

Key Capabilities

  • Enhanced Reasoning: Optimized to improve Chain-of-Thought (CoT) reasoning, leading to more logical and coherent outputs.
  • Structured Response Quality: Specifically trained to produce higher quality and better-structured responses based on preferred outputs.
  • Direct Use: As a fully merged model, it can be loaded and used directly with the transformers library, simplifying deployment.

Training Details

The model was trained for 1 epoch with a learning rate of 1e-07 and a beta value of 0.1, using a maximum sequence length of 1024. The training utilized the u-10bei/dpo-dataset-qwen-cot dataset, focusing on preference alignment for reasoning tasks.

Good For

  • Applications requiring improved logical reasoning and problem-solving.
  • Generating structured and high-quality text responses.
  • Developers looking for a Qwen3-4B variant with enhanced CoT and response quality.