Human-Like-Qwen2.5-1.5B-Instruct on Ollama

Human-Like-Qwen2.5-1.5B-Instruct is a fine-tuned language model based on Qwen2.5-1.5B-Instruct, optimized using the Human-Like-DPO-Dataset. This model is designed for more natural, engaging, and human-like conversations.

This guide provides steps to install, run, and customize the model on Ollama.

🚀 Features

Fine-tuned from Qwen2.5-1.5B-Instruct
Improved conversational fluency and engagement
Enhanced response ranking using DPO (Direct Preference Optimization)
Context length up to 32K tokens
Supports multi-turn conversations with coherent, human-like responses

📥 Installation

1️⃣ Install Ollama

If you haven’t installed Ollama, follow these steps:

For macOS & Linux:

curl -fsSL https://ollama.com/install.sh | sh

For Windows (WSL required):

wsl --install
curl -fsSL https://ollama.com/install.sh | sh

For more details, check the official Ollama installation guide:
https://ollama.com/download

🔧 Running the Model

2️⃣ Pull the Model

After installing Ollama, download Human-Like-Qwen2.5-1.5B-Instruct:

ollama pull myrepo/human-like-qwen2.5-1.5b-instruct

3️⃣ Run the Model

To start generating responses:

ollama run myrepo/human-like-qwen2.5-1.5b-instruct

For interactive chat mode:

ollama chat myrepo/human-like-qwen2.5-1.5b-instruct

📌 Customizing the Model

Using a Custom `Modelfile`

You can modify the model behavior using a Modelfile.

Create a new file named Modelfile and add the following:

FROM myrepo/human-like-qwen2.5-1.5b-instruct

PARAMETER temperature 0.7
PARAMETER top_p 0.9

SYSTEM "You are a human-like AI assistant, capable of engaging conversations, providing informative answers, and assisting with a variety of topics in a natural and friendly manner."

Build and run your custom model:

ollama create human-like-qwen-custom -f Modelfile
ollama run human-like-qwen-custom

📄 License

This model is released under the MIT License, ensuring open and accessible AI development.

🔗 Resources

Hugging Face Repository: (if applicable)
Ollama Documentation: https://ollama.com/docs
GitHub Repository: (if applicable)

🙌 Acknowledgments

This model was fine-tuned using the Human-Like-DPO-Dataset to improve conversational fluency and human-like interactions. Special thanks to the developers of Qwen2.5-1.5B-Instruct and the HumanLLMs team for their contributions to DPO-based fine-tuning.

Readme