Edit model card

T5-base fine-tuned on CommonGen

Google's T5 fine-tuned on CommonGen for Generative Commonsense Reasoning.

Details of T5

The T5 model was presented in Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer by Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu in Here the abstract:

Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new β€œColossal Clean Crawled Corpus”, we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.

model image

Details of the dataset πŸ“š

CommonGen is a constrained text generation task, associated with a benchmark dataset, to explicitly test machines for the ability of generative commonsense reasoning. Given a set of common concepts; the task is to generate a coherent sentence describing an everyday scenario using these concepts.

CommonGen is challenging because it inherently requires 1) relational reasoning using background commonsense knowledge, and 2) compositional generalization ability to work on unseen concept combinations. Our dataset, constructed through a combination of crowd-sourcing from AMT and existing caption corpora, consists of 30k concept-sets and 50k sentences in total.

Dataset Split # samples
common_gen train 67389
common_gen valid 4018
common_gen test 1497

Model fine-tuning πŸ‹οΈβ€

The training script is a slightly modified version of this awesome one by Suraj Patil

Metrics πŸ“‹

Metric Score
ROUGE-2 17.10
ROUGE-L 39.47
BLEU WIP

The metrics above slightly improves results shown in the paper for the same model and metrics.

Model in Action πŸš€

from transformers import AutoModelWithLMHead, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained("mrm8488/t5-base-finetuned-common_gen")
model = AutoModelWithLMHead.from_pretrained("mrm8488/t5-base-finetuned-common_gen")

def gen_sentence(words, max_length=32):
  input_text = words
  features = tokenizer([input_text], return_tensors='pt')

  output = model.generate(input_ids=features['input_ids'], 
               attention_mask=features['attention_mask'],
               max_length=max_length)

  return tokenizer.decode(output[0], skip_special_tokens=True)

words = "tree plant ground hole dig"

gen_sentence(words)

# output: digging a hole in the ground to plant trees

Open In Colab

Created by Manuel Romero/@mrm8488 | LinkedIn

Made with β™₯ in Spain

Downloads last month
957
Safetensors
Model size
297M params
Tensor type
F32
Β·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for mrm8488/t5-base-finetuned-common_gen

Finetunes
4 models

Dataset used to train mrm8488/t5-base-finetuned-common_gen

Spaces using mrm8488/t5-base-finetuned-common_gen 3