Overview

This model, dpo-qwen-cot-merged, is a 4 billion parameter language model based on the Qwen3 architecture, specifically fine-tuned from Qwen/Qwen3-4B-Instruct-2507. Developed by takeshi200ok, it leverages Direct Preference Optimization (DPO) via the Unsloth library to align its responses with preferred outputs.

Key Capabilities

• Enhanced Reasoning: Optimized to improve Chain-of-Thought (CoT) reasoning, making it suitable for tasks requiring multi-step logical deduction.
• Improved Structured Responses: DPO training focuses on generating higher quality and more structured outputs.
• Fully Merged Model: Provided as a complete 16-bit model, eliminating the need for separate adapter loading.

Training Details

The model underwent 1 epoch of DPO training with a learning rate of 1e-05 and a beta value of 0.1. It utilized a maximum sequence length of 3072 tokens. The training data was sourced from u-10bei/dpo-dataset-qwen-cot.

Good For

• Applications requiring robust reasoning and logical inference.
• Generating structured and coherent text outputs.
• Use cases where response alignment and quality are critical.