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 downstream task (QDMRs) - Dataset 📚
Break is a human annotated dataset of natural language questions and their Question Decomposition Meaning Representations (QDMRs). Break consists of 83,978 examples sampled from 10 question answering datasets over text, images and databases. This repository contains the Break dataset along with information on the exact data format.
Dataset
Split
# samples
break_data
train
17503
break_data
valid
3130
Check out more about this dataset and others in
NLP Viewer
Model fine-tuning 🏋️
The training script is a slightly modified version of
this awesome one
by
Suraj Patil
. The main change is at preprocessing
inputs
and
targets
we feed to the model. We do it as a
paraphrasing task
.
Model in Action 🚀
# Tip: By now, install transformers from sourcefrom transformers import AutoModelForSeq2SeqLM, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("mrm8488/t5-base-finetuned-break_data")
model = AutoModelForSeq2SeqLM.from_pretrained("mrm8488/t5-base-finetuned-break_data")
defget_decomposition(question):
input_text = "paraphrase: %s </s>" % question
features = tokenizer([input_text], return_tensors='pt')
output = model.generate(input_ids=features['input_ids'],
attention_mask=features['attention_mask'],
max_length=32)
return tokenizer.decode(output[0])
question = "The composer of Sands Theme plays what type of guitar?"
get_decomposition(question)
# output: 'return Sands Theme ;return composer of #1 ;return guitar that #2 plays'
Runs of mrm8488 t5-base-finetuned-break_data on huggingface.co
69
Total runs
-2
24-hour runs
4
3-day runs
14
7-day runs
53
30-day runs
More Information About t5-base-finetuned-break_data huggingface.co Model
t5-base-finetuned-break_data huggingface.co
t5-base-finetuned-break_data huggingface.co is an AI model on huggingface.co that provides t5-base-finetuned-break_data's model effect (), which can be used instantly with this mrm8488 t5-base-finetuned-break_data model. huggingface.co supports a free trial of the t5-base-finetuned-break_data model, and also provides paid use of the t5-base-finetuned-break_data. Support call t5-base-finetuned-break_data model through api, including Node.js, Python, http.
t5-base-finetuned-break_data huggingface.co is an online trial and call api platform, which integrates t5-base-finetuned-break_data's modeling effects, including api services, and provides a free online trial of t5-base-finetuned-break_data, you can try t5-base-finetuned-break_data online for free by clicking the link below.
mrm8488 t5-base-finetuned-break_data online free url in huggingface.co:
t5-base-finetuned-break_data is an open source model from GitHub that offers a free installation service, and any user can find t5-base-finetuned-break_data on GitHub to install. At the same time, huggingface.co provides the effect of t5-base-finetuned-break_data install, users can directly use t5-base-finetuned-break_data installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
t5-base-finetuned-break_data install url in huggingface.co: