Spaces:
Running
on
Zero
Running
on
Zero
| title: Psychology AI Assistant | |
| emoji: 🧠 | |
| colorFrom: purple | |
| colorTo: blue | |
| sdk: gradio | |
| sdk_version: 5.34.2 | |
| app_file: app.py | |
| pinned: false | |
| license: apache-2.0 | |
| models: | |
| - Qwen/Qwen2.5-0.5B | |
| - phxdev/psychology-qwen-0.5b | |
| tags: | |
| - psychology | |
| - mental-health | |
| - therapy | |
| - counseling | |
| - peft | |
| - qwen | |
| - fine-tuned | |
| # 🧠 Psychology AI Assistant | |
| An empathetic AI assistant fine-tuned for psychology and mental health support. Based on Qwen 2.5-0.5B with specialized training using PEFT/LoRA techniques. | |
| ## ✨ Features | |
| - **Empathetic Responses**: Fine-tuned specifically for psychology and mental health conversations | |
| - **Multiple Support Styles**: Choose from therapeutic, supportive friend, crisis support, anxiety coaching, or mindfulness guidance | |
| - **Evidence-Based Approach**: Responses grounded in psychological principles and therapeutic techniques | |
| - **Interactive Interface**: Modern, accessible design with conversation starters and customizable settings | |
| - **Privacy-Focused**: No conversation data is stored or logged | |
| ## 🎯 Model Details | |
| - **Base Model**: [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) | |
| - **Fine-tuned Model**: [phxdev/psychology-qwen-0.5b](https://huggingface.co/phxdev/psychology-qwen-0.5b) | |
| - **Training Method**: PEFT (Parameter Efficient Fine-Tuning) with LoRA | |
| - **Specialization**: Psychology, mental health, therapeutic communication | |
| - **Parameters**: ~500M parameters (0.5B) | |
| - **Context Length**: 2048 tokens | |
| ## 🚀 How to Use | |
| 1. **Start a Conversation**: Type your thoughts, concerns, or questions in the chat box | |
| 2. **Choose Your Style**: Select the type of support you prefer from the dropdown menu | |
| 3. **Adjust Settings**: Customize response creativity and length to your preference | |
| 4. **Try Examples**: Click on any of the conversation starters for common scenarios | |
| ### Support Styles Available: | |
| - **🔬 Therapeutic**: Professional therapist approach with validation and structured guidance | |
| - **👥 Supportive Friend**: Warm, caring friend who studied psychology | |
| - **🆘 Crisis Support**: Immediate support and grounding techniques for crisis situations | |
| - **😰 Anxiety Coach**: Specialized techniques for managing anxiety and panic | |
| - **🧘 Mindfulness Guide**: Present-moment awareness and self-compassion practices | |
| ## ⚠️ Important Disclaimer | |
| This AI assistant provides supportive guidance based on psychological principles, but it is **not a replacement for professional therapy or medical advice**. | |
| If you're experiencing a mental health crisis, please contact: | |
| - **988 Suicide & Crisis Lifeline**: Call or text 988 | |
| - **Crisis Text Line**: Text HOME to 741741 | |
| - **Emergency Services**: Call 911 (US) or your local emergency number | |
| ## 🛠️ Technical Implementation | |
| The model uses advanced prompt engineering and fine-tuning techniques to provide contextually appropriate responses: | |
| - **PEFT Training**: Efficient fine-tuning preserving base model capabilities while adding psychology specialization | |
| - **Prompt Templates**: Multiple specialized templates for different therapeutic approaches | |
| - **Response Quality**: Built-in safeguards against generic or unhelpful responses | |
| - **Mobile Optimized**: Responsive design works well on all devices | |
| ## 🎨 Interface Features | |
| - **Clean Design**: Modern, calming interface designed for sensitive conversations | |
| - **Conversation History**: Full chat history maintained during your session | |
| - **Example Prompts**: Curated conversation starters for common mental health topics | |
| - **Accessibility**: Screen reader friendly and keyboard navigable | |
| - **Privacy**: No data logging or conversation storage | |
| ## 🔧 Model Performance | |
| The fine-tuned model shows significant improvements over the base Qwen model in: | |
| - **Empathy**: Better emotional validation and understanding | |
| - **Practical Advice**: More actionable, specific guidance | |
| - **Psychological Awareness**: Understanding of mental health concepts and terminology | |
| - **Conversational Flow**: Natural, supportive dialogue without generic responses | |
| ## 📊 Example Use Cases | |
| - **Stress Management**: Work stress, academic pressure, life transitions | |
| - **Anxiety Support**: Panic attacks, social anxiety, general worry | |
| - **Relationship Guidance**: Communication skills, boundary setting, conflict resolution | |
| - **Self-Esteem**: Building confidence, overcoming negative self-talk | |
| - **Mindfulness**: Present-moment awareness, meditation guidance, self-compassion | |
| - **Grief Support**: Processing loss, coping with change | |
| ## 🤝 Contributing & Feedback | |
| This model is part of ongoing research into AI applications for mental health support. If you have feedback or suggestions: | |
| - **Model Issues**: Report on the [model repository](https://huggingface.co/phxdev/psychology-qwen-0.5b) | |
| - **Interface Bugs**: Open an issue on this Space | |
| - **Research Collaboration**: Contact [@phxdev](https://huggingface.co/phxdev) | |
| ## 📚 Related Resources | |
| - **Base Model**: [Qwen/Qwen2.5-0.5B](https://huggingface.co/Qwen/Qwen2.5-0.5B) | |
| - **PEFT Library**: [huggingface/peft](https://github.com/huggingface/peft) | |
| - **Mental Health Resources**: [National Alliance on Mental Illness](https://nami.org) | |
| ## 📄 License | |
| This Space and the fine-tuned model are released under the Apache 2.0 License. The base Qwen model follows its original licensing terms. | |
| --- | |
| **Built with ❤️ by [@phxdev](https://huggingface.co/phxdev) using Gradio and HuggingFace Transformers** |