Health AI | Emergent Ventures Fellow | 1517 Fund Grantee | Intern at Dartmouth Hitchcock Center | TKS Alumni
Hey! I'm building a chatbot over the next 20 days that can provide personalized recommendations for cardiovascular conditions and features multilingual and voice capabilities. ---Day 8--- I had to redo my data.json file from yesterday because I realized there are 2 ways I can get better responses from my model: 1. Creating a vector database using Pinecone 2. Finetuning I chose to go down the route of fine-tuning, but I had to reformat my data.json file to match the OpenAI API documentation: https://lnkd.in/epmkRFkz Here's a great video explaining the fine-tuning documentation step by step: https://lnkd.in/eVJDRnNg Thank you Aleem Rehmtulla for helping to build this out! I would really appreciate any help to build up the 2nd version - would love to connect with experts in this space! PS: I got chatgpt to write my data.json file #promptengineering