gsarti/it5-small

Italian T5 Small 🇮🇹

The IT5 model family represents the first effort in pretraining large-scale sequence-to-sequence transformer models for the Italian language, following the approach adopted by the original T5 model.

This model is released as part of the project "IT5: Text-to-Text Pretraining for Italian Language Understanding and Generation", by Gabriele Sarti and Malvina Nissim with the support of Huggingface and with TPU usage sponsored by Google's TPU Research Cloud. All the training was conducted on a single TPU3v8-VM machine on Google Cloud. Refer to the Tensorboard tab of the repository for an overview of the training process.

The inference widget is deactivated because the model needs a task-specific seq2seq fine-tuning on a downstream task to be useful in practice. The models in the it5 organization provide some examples of this model fine-tuned on various downstream task.

Model variants

This repository contains the checkpoints for the base version of the model. The model was trained for one epoch (1.05M steps) on the Thoroughly Cleaned Italian mC4 Corpus (~41B words, ~275GB) using 🤗 Datasets and the google/t5-v1_1-small improved configuration. The training procedure is made available on Github.

The following table summarizes the parameters for all available models

	it5-small (this one)	it5-base	it5-large	it5-base-oscar
dataset	gsarti/clean_mc4_it	gsarti/clean_mc4_it	gsarti/clean_mc4_it	oscar/unshuffled_deduplicated_it
architecture	google/t5-v1_1-small	google/t5-v1_1-base	google/t5-v1_1-large	t5-base
learning rate	5e-3	5e-3	5e-3	1e-2
steps	1'050'000	1'050'000	2'100'000	258'000
training time	36 hours	101 hours	370 hours	98 hours
ff projection	gated-gelu	gated-gelu	gated-gelu	relu
tie embeds	false	false	false	true
optimizer	adafactor	adafactor	adafactor	adafactor
max seq. length	512	512	512	512
per-device batch size	16	16	8	16
tot. batch size	128	128	64	128
weigth decay	1e-3	1e-3	1e-2	1e-3
validation split size	15K examples	15K examples	15K examples	15K examples

The high training time of it5-base-oscar was due to a bug in the training script.

For a list of individual model parameters, refer to the config.json file in the respective repositories.

Using the models

from transformers import AutoTokenzier, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("gsarti/it5-small")
model = AutoModelForSeq2SeqLM.from_pretrained("gsarti/it5-small")

Note: You will need to fine-tune the model on your downstream seq2seq task to use it. See an example here.

Flax and Tensorflow versions of the model are also available:

from transformers import FlaxT5ForConditionalGeneration, TFT5ForConditionalGeneration

model_flax = FlaxT5ForConditionalGeneration.from_pretrained("gsarti/it5-small")
model_tf = TFT5ForConditionalGeneration.from_pretrained("gsarti/it5-small")

Limitations

Due to the nature of the web-scraped corpus on which IT5 models were trained, it is likely that their usage could reproduce and amplify pre-existing biases in the data, resulting in potentially harmful content such as racial or gender stereotypes and conspiracist views. For this reason, the study of such biases is explicitly encouraged, and model usage should ideally be restricted to research-oriented and non-user-facing endeavors.

Model curators

For problems or updates on this model, please contact gabriele.sarti996@gmail.com.

Citation Information

@inproceedings{sarti-nissim-2024-it5-text,
    title = "{IT}5: Text-to-text Pretraining for {I}talian Language Understanding and Generation",
    author = "Sarti, Gabriele  and
      Nissim, Malvina",
    editor = "Calzolari, Nicoletta  and
      Kan, Min-Yen  and
      Hoste, Veronique  and
      Lenci, Alessandro  and
      Sakti, Sakriani  and
      Xue, Nianwen",
    booktitle = "Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)",
    month = may,
    year = "2024",
    address = "Torino, Italia",
    publisher = "ELRA and ICCL",
    url = "https://aclanthology.org/2024.lrec-main.823",
    pages = "9422--9433",
}