Closed quantumalchemy closed 10 months ago
@quantumalchemy , if you have any pointers on how you did this, it would be much appreciated!
@quantumalchemy Thanks for your kind words!
@AntonOfTheWoods I think @quantumalchemy is fine-tuning LLM on QA, using QG model generation via LMQG, but not fine-tuning LLMs on QG.
Wish I could have found this earlier.. Works great! saves time & $$ dont have to use openai api ( as its very costly if you have large datasets) Using it to create q&a alpaca datasets for 7b LLM these small llms / training: respond very well with this type of data - Just Great!