README.md

from datasets import load_dataset from trl import SFTTrainer

dataset = load_dataset("imdb", split="train")

trainer = SFTTrainer( "facebook/opt-350m", train_dataset=dataset, dataset_text_field="text", max_seq_length=512, ) trainer.train()