ajsbsd/

presto

Code to test on Colab !pip install -q transformers[torch] tokenizers datasets evaluate rouge_score sentencepiece huggingface_hub --upgrade

from huggingface_hub import notebook_login

notebook_login()

import nltk from datasets import load_dataset import evaluate import numpy as np from transformers import T5Tokenizer, DataCollatorForSeq2Seq from transformers import T5ForConditionalGeneration, Seq2SeqTrainingArguments, Seq2SeqTrainer

dataset =… See the full description on the dataset page: https://huggingface.co/datasets/ajsbsd/presto.