Model Card: Fine-tuned Phi3-mini for Mental Health FAQs
Model Details
- Model Name: Phi3-mini-MentalHealth
- Base Model: Phi3-mini
- Model Type: Language Model
- Framework: MLX
- Fine-tuning Dataset: Mental Health FAQ data from Kaggle
- Format: GGUF (GPT Quantized)
Intended Use
This model is designed to provide empathetic and informative responses to frequently asked questions about mental health. It can be used in applications such as:
- Mental health chatbots
- Educational resources about mental health
- Preliminary mental health screening tools (with appropriate disclaimers)
Training Data
The model was fine-tuned on a dataset of Mental Health FAQs from Kaggle. This dataset likely includes:
- Common questions about mental health conditions
- Basic information about symptoms and treatments
- General mental health and wellness advice
Note: The exact contents and size of the dataset should be verified and added here.
Fine-tuning Process
The Phi3-mini model was fine-tuned using the MLX framework. Specific details about the fine-tuning process, such as number of epochs, learning rate, and other hyperparameters, should be added here.
Performance and Limitations
- Improved Empathy: The fine-tuned model demonstrates increased empathy in its responses compared to the base model.
- Domain Specificity: The model's knowledge is focused on mental health topics covered in the FAQ dataset.
- Limitations:
- The model's knowledge is limited to the scope of the training data.
- It should not be used as a substitute for professional mental health advice or diagnosis.
- The model may not be up-to-date with the latest mental health research or treatments.
Ethical Considerations
- Privacy: The model should be used in ways that protect user privacy, especially given the sensitive nature of mental health discussions.
- Bias: While efforts have been made to improve empathy, the model may still exhibit biases present in the training data.
- Misuse Potential: Care should be taken to prevent the model from being used to provide medical advice it's not qualified to give.
Additional Information
- License: Apache 2.0
- Developer: Dattaraj Rao - Persistent
- Date Created: 16-Jul-2024
- Version: 1.0
For more information or to report issues, please contact:
https://www.linkedin.com/in/dattarajrao/
