Spaces:
Running
Running
Finetuned a Early 2023-Era Model on 2 Instruction Following Datasets and it Became Good
#9
by Tralalabs - opened
Well, I finetuned Pythia-6.9B for 550 steps for Instruction Following and it became good. The raw, base model didn't know non-English languages. It knowed a little, but....... the finetuned one knows 13 languages! Even with loops, the model IS GOOD. https://huggingface.co/Tralalabs/Pythia-6.9B-Instruct-v1-Merged
LH-Tech-AI changed discussion status to closed