X-D-Lab/MindChat-Qwen2-4B

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:Feb 4, 2024License:otherArchitecture:Transformer0.0K Warm

MindChat-Qwen2-4B is a 4 billion parameter psychological large language model developed by X-D-Lab, built upon the Qwen2 architecture with a 32768 token context length. It is specifically fine-tuned on approximately 200,000 high-quality multi-turn psychological dialogue data, covering various aspects like work, family, and social interactions. This model is designed to provide psychological comfort, assessment, diagnosis, and support, aiming to alleviate mental stress and address psychological concerns while strictly protecting user privacy.

Loading preview...

MindChat-Qwen2-4B: A Specialized Psychological LLM

MindChat-Qwen2-4B, developed by X-D-Lab, is a 4 billion parameter large language model built on the Qwen2 architecture, featuring a 32768 token context length. It is part of the MindChat series, which focuses on mental health support. This model is specifically trained on a high-quality dataset of approximately 200,000 multi-turn psychological dialogues, encompassing diverse topics such as work, family, learning, and social relationships.

Key Capabilities

  • Psychological Support: Designed to assist users with psychological consultation, assessment, diagnosis, and therapy-like interactions.
  • Privacy-Focused: Emphasizes creating a private, warm, safe, timely, and convenient dialogue environment.
  • Empathy and Alignment: Aims to provide empathetic responses and align with human values, as highlighted in earlier MindChat versions.
  • Accessibility: The 4B parameter size allows for deployment in various scenarios, including personal PCs or mobile devices, enhancing user privacy.

Good for

  • Mental Health Applications: Ideal for developing applications focused on psychological comfort and preliminary mental health support.
  • Research in Computational Psychology: Useful for academic research into large language models for psychological domains.
  • Privacy-Sensitive Dialogue Systems: Suitable for use cases where user privacy in sensitive conversations is paramount.
  • Educational Tools: Can be adapted for educational purposes to help users understand and manage psychological well-being.