a080fe0 5dcd105 a080fe0
1
2
3
4
5
6
7
8
9
10
from trl import SFTTrainer from datasets import load_dataset dataset = load_dataset("trl-lib/Capybara", split="train") trainer = SFTTrainer( model="openai-community/gpt2", train_dataset=dataset, ) trainer.train()
