joaHofmann/Qwen2.5-0.5B-DPO

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 12, 2026Architecture:Transformer Featherless Exclusive Cold

The joaHofmann/Qwen2.5-0.5B-DPO is a 0.5 billion parameter language model, fine-tuned from Qwen/Qwen2.5-0.5B-Instruct using Direct Preference Optimization (DPO). This model is designed to align its outputs more closely with human preferences, leveraging the DPO method for improved response quality. It is suitable for conversational AI and instruction-following tasks where nuanced and preferred responses are critical.

Loading preview...

Model Overview

The joaHofmann/Qwen2.5-0.5B-DPO is a 0.5 billion parameter language model derived from the Qwen/Qwen2.5-0.5B-Instruct base model. It has been specifically fine-tuned using Direct Preference Optimization (DPO), a method that aligns language model outputs with human preferences by treating the language model as a reward model.

Key Characteristics

  • Base Model: Fine-tuned from Qwen/Qwen2.5-0.5B-Instruct.
  • Parameter Count: 0.5 billion parameters.
  • Context Length: Supports a context length of 32768 tokens.
  • Training Method: Utilizes Direct Preference Optimization (DPO) for enhanced alignment with preferred responses.
  • Framework: Training was conducted using the TRL library.

Use Cases

This model is particularly well-suited for applications requiring:

  • Instruction Following: Generating responses that adhere to specific user instructions.
  • Conversational AI: Producing more natural and preferred dialogue in chatbots and virtual assistants.
  • Preference Alignment: Scenarios where the quality and style of generated text need to match human preferences effectively.