Hello everyone, I am currently fine tuning GPT-3 and looking at various metrics to see if the training process has gone sufficiently well. Now I am wondering if it is possible to fine tune fine tuned models. The idea is that: feed GPT-3 with knowledge about a specific topic (e.g. reforestation …

Fine tune fine tuned models

knickerm April 6, 2022, 3:03pm 3

yes sure, I have a big data set of prompt/completion examples.

1 Like