File size: 1,041 Bytes

97d5a09

Steps taken:
1 - I extracted one disease from the pdf. I realized all the diseases have one template format. I extracted that into Gpt4 and then added just one disease. (This needs to be done for all the diseases or information, we have to figure out a way to represent it.)
2 - After adding one diseases i asked codeinterpreter to create a json for me, and then flatten that json so pandas can read it.
3 - I asked Gpt4 to create for me a synthetic dataset based on the format I gave it. I landed on this article: https://www.datacamp.com/tutorial/fine-tuning-gpt-3-using-the-open-ai-api-and-python
The article made things easy for me, using one-shot learning Gpt4 was able to create a dataset for me under simple instructions.
4 - Now I just need to set up with the API key and test on the API

Another useful link: https://medium.com/@_sumitsaha_/fine-tuning-openais-gpt3-model-fd9cb06517f6#:~:text=Fine%2Dtuning%20allows%20the%20model,efficient%2C%20and%20easier%20to%20train.

Note: This file contains experimentation by Atwine Mostly.