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 “Colossal 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.
Details of the dataset 📚
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.
The metrics above slightly improves results shown in the
paper
for the same model and metrics.
Model in Action 🚀
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")
defgen_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
Runs of mrm8488 t5-base-finetuned-common_gen on huggingface.co
310
Total runs
0
24-hour runs
12
3-day runs
47
7-day runs
-326
30-day runs
More Information About t5-base-finetuned-common_gen huggingface.co Model
t5-base-finetuned-common_gen huggingface.co
t5-base-finetuned-common_gen huggingface.co is an AI model on huggingface.co that provides t5-base-finetuned-common_gen's model effect (), which can be used instantly with this mrm8488 t5-base-finetuned-common_gen model. huggingface.co supports a free trial of the t5-base-finetuned-common_gen model, and also provides paid use of the t5-base-finetuned-common_gen. Support call t5-base-finetuned-common_gen model through api, including Node.js, Python, http.
t5-base-finetuned-common_gen huggingface.co is an online trial and call api platform, which integrates t5-base-finetuned-common_gen's modeling effects, including api services, and provides a free online trial of t5-base-finetuned-common_gen, you can try t5-base-finetuned-common_gen online for free by clicking the link below.
mrm8488 t5-base-finetuned-common_gen online free url in huggingface.co:
t5-base-finetuned-common_gen is an open source model from GitHub that offers a free installation service, and any user can find t5-base-finetuned-common_gen on GitHub to install. At the same time, huggingface.co provides the effect of t5-base-finetuned-common_gen install, users can directly use t5-base-finetuned-common_gen installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
t5-base-finetuned-common_gen install url in huggingface.co: