Diocletianus/dpo-qwen-cot-merged0207 is a 4 billion parameter language model fine-tuned from Qwen/Qwen3-4B-Instruct-2507 using Direct Preference Optimization (DPO) via Unsloth. This model is specifically optimized to improve reasoning capabilities, particularly Chain-of-Thought (CoT), and enhance structured response quality. It is designed for applications requiring aligned and coherent outputs based on preferred response patterns.
Loading preview...
Model Overview
Diocletianus/dpo-qwen-cot-merged0207 is a 4 billion parameter language model derived from the Qwen3-4B-Instruct-2507 base model. It has been fine-tuned using Direct Preference Optimization (DPO) with the Unsloth library, resulting in a merged 16-bit weight model that requires no adapter loading.
Key Capabilities
- Enhanced Reasoning: Optimized through DPO to improve Chain-of-Thought (CoT) reasoning, enabling more structured and logical problem-solving.
- Aligned Responses: Fine-tuned to align its outputs with preferred response patterns, leading to higher quality and more relevant generations.
- Structured Output: Focuses on improving the quality of structured responses, making it suitable for tasks requiring specific formats or coherent arguments.
Training Details
The model underwent 1 epoch of DPO training with a learning rate of 1e-07 and a beta value of 0.1. The maximum sequence length used during training was 1024 tokens. The training utilized the u-10bei/dpo-dataset-qwen-cot dataset. The LoRA configuration (r=8, alpha=16) was merged directly into the base model.
Usage
This merged model can be directly integrated and used with the transformers library for inference, providing a straightforward deployment experience. Users should adhere to the MIT License of the training data and the original base model's license terms.