README.md

metadata

title: Phi-2 Fine-tuned Assistant
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 4.0.0
app_file: app.py
pinned: false

Phi-2 Fine-tuned Assistant

This is a fine-tuned version of Microsoft's Phi-2 model on the OpenAssistant dataset. The model has been trained to follow instructions and provide helpful responses.

Usage

Simply enter your instruction in the text box and click submit. The model will generate a response based on your input.

Examples

• "Explain what is Python in one sentence."
• "How do neural networks learn through backpropagation?"
• "Write a short poem about coding."

Model Details

• Base Model: Microsoft Phi-2
• Training Data: OpenAssistant dataset
• Fine-tuning Method: QLoRA
• Deployment: CPU inference

Limitations

• The model runs on CPU, so responses might take a few seconds
• Maximum input length is 256 tokens
• Maximum output length is 128 tokens