Finetuning tutorial for absolute noobs

I could have sworn I’d posted this. The first two videos are up. The third will be how to actually use a finetuned model.

Stay tuned! I was given an interesting challenge by a member of my discord server: using GPT-3 to match resumes to job descriptions. I’ll probably get that experiment posted tomorrow. In the meantime: finetuning tutorial!


I tried to watch this but couldn’t as it said video was private

Hi Dave, I have a really newbie question related to fine tuning the gpt3 thru api command. I was trying to use your “question generator” jsonl file as trining data to fine tune the ada model. It kept giving me this error saying “The OpenAI API can only be accessed over HTTPS. You should access rather than the current URL. (HTTP status code: 403)”.

I’ve tested the endpoint using python request lib, and got a 404 result. Note that I’m pyhsically in China so I have to use VPN service to access the openai api. How can i fix this issue so that the fine tune command can be executed. Thanks a lot!

PS: I’ve interacted with ChatGPT with this issue and understand this is casued by my system is not allowing https traffic in/out the api endpoint. I should somehow unblock the endpoint or allowing the https traffic. However, I just can’t figure out how to do it.