Data

Tools

Indexes

Applications

Experiments

Open Source Science

tasksource

Datasets collection and preprocessings framework for NLP extreme multitask learning

https://github.com/sileod/tasksource

Science Score: 64.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ✓
    CITATION.cff file
    Found CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ✓
    Academic publication links
    Links to: arxiv.org
  • ✓
    Committers with academic emails
    1 of 2 committers (50.0%) from academic institutions
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (12.2%) to scientific vocabulary

Keywords

benchmark bigbench crossfit curated-datasets dataset-collection discriminative extreme-mtl extreme-multi-task-learning glue huggingface instruction-tuning meta-learning multi-task-learning multi-task-learning-scaling natural-language-inference nlp preprocessings reward-modeling scaling text-classification
Last synced: 6 months ago · JSON representation ·

Repository

Datasets collection and preprocessings framework for NLP extreme multitask learning

Basic Info
  • Host: GitHub
  • Owner: sileod
  • License: cc-by-4.0
  • Language: Python
  • Default Branch: main
  • Homepage:
  • Size: 376 KB
Statistics
  • Stars: 184
  • Watchers: 3
  • Forks: 10
  • Open Issues: 4
  • Releases: 48
Topics
benchmark bigbench crossfit curated-datasets dataset-collection discriminative extreme-mtl extreme-multi-task-learning glue huggingface instruction-tuning meta-learning multi-task-learning multi-task-learning-scaling natural-language-inference nlp preprocessings reward-modeling scaling text-classification
Created about 3 years ago · Last pushed 8 months ago
Metadata Files
Readme License Citation

README.md

tasksource 600+ curated datasets and preprocessings for instant and interchangeable use

Huggingface Datasets is an excellent library, but it lacks standardization, and datasets often require preprocessing work to be used interchangeably. tasksource streamlines interchangeable datasets usage to scale evaluation or multi-task learning.

Each dataset is standardized to a MultipleChoice, Classification, or TokenClassification template with canonical fields. We focus on discriminative tasks (= with negative examples or classes) for our annotations but also provide a SequenceToSequence template. All implemented preprocessings are in tasks.py or tasks.md. A preprocessing is a function that accepts a dataset and returns the standardized dataset. Preprocessing code is concise and human-readable.

Installation and usage:

pip install tasksource ```python from tasksource import listtasks, loadtask df = list_tasks(multilingual=False) # takes some time

for id in df[df.tasktype=="MultipleChoice"].id: dataset = loadtask(id) # all yielded datasets can be used interchangeably ```

Browse the 500+ curated tasks in tasks.md (200+ MultipleChoice tasks, 200+ Classification tasks), and feel free to request a new task. Datasets are downloaded to $HF_DATASETS_CACHE (like any Hugging Face dataset), so ensure you have more than 100GB of space available.

You can now also use: python load_dataset("tasksource/data", "glue/rte",max_rows=30_000)

Pretrained models:

Text encoder pretrained on tasksource reached state-of-the-art results: 🤗/deberta-v3-base-tasksource-nli

Tasksource pretraining is notably helpful for RLHF reward modeling or any kind of classification, including zero-shot. You can also find a large and a multilingual version.

tasksource-instruct

The repo also contains some recasting code to convert tasksource datasets to instructions, providing one of the richest instruction-tuning datasets: 🤗/tasksource-instruct-v0

tasksource-label-nli

We also recast all classification tasks as natural language inference, to improve entailment-based zero-shot classification detection: 🤗/zero-shot-label-nli

Write and use custom preprocessings

```python from tasksource import MultipleChoice

codah = MultipleChoice('questionpropmt',choiceslist='candidateanswers', labels='correctansweridx', datasetname='codah', config_name='codah')

winogrande = MultipleChoice('sentence',['option1','option2'],'answer', datasetname='winogrande',configname='winogrande_xl', splits=['train','validation',None]) # test labels are not usable

tasks = [winogrande.load(), codah.load()]) # Aligned datasets (same columns) can be used interchangably
```

### Citation and contact

For more details, refer to this article: bib @inproceedings{sileo-2024-tasksource, title = "tasksource: A Large Collection of {NLP} tasks with a Structured Dataset Preprocessing Framework", author = "Sileo, Damien", 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.1361", pages = "15655--15684", } For help integrating tasksource into your experiments, please contact damien.sileo@inria.fr.

Owner

  • Login: sileod
  • Kind: user

Damien Sileo

Citation (CITATION.cff) 

cff-version: 1.1.0
message: "If you use this work, please cite it as below."
authors:
  - family-names: "Sileo"
    given-names: "Damien"
title: "tasksource: A Dataset Harmonization Framework for Streamlined NLP Multi-Task Learning and Evaluation"
version: "1.0.0"
date-released: 2023-01-01
url: "https://arxiv.org/abs/2301.05948"

GitHub Events

Total
  • Release event: 1
  • Watch event: 39
  • Issue comment event: 1
  • Push event: 3
  • Pull request event: 2
  • Fork event: 2
  • Create event: 2
Last Year
  • Release event: 1
  • Watch event: 39
  • Issue comment event: 1
  • Push event: 3
  • Pull request event: 2
  • Fork event: 2
  • Create event: 2

Committers

Last synced: almost 3 years ago

All Time
  • Total Commits: 132
  • Total Committers: 2
  • Avg Commits per committer: 66.0
  • Development Distribution Score (DDS): 0.038
Top Committers
Name Email Commits
sileod d****o@g****m 127
Damien Sileo d****o@m****r 5
Committer Domains (Top 20 + Academic)
magnet-magnet7.lille.inria.fr: 1

Issues and Pull Requests

Last synced: 8 months ago

All Time
  • Total issues: 9
  • Total pull requests: 0
  • Average time to close issues: about 1 month
  • Average time to close pull requests: N/A
  • Total issue authors: 9
  • Total pull request authors: 0
  • Average comments per issue: 1.44
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
View more stats
Top Authors
Issue Authors
  • cuongnguyengit (1)
  • pesvut (1)
  • avidale (1)
  • assaftibm (1)
  • Deehan1866 (1)
  • A1exRey (1)
  • nickypro (1)
  • nivibilla (1)
  • imoneoi (1)
Pull Request Authors
Top Labels
Issue Labels
enhancement (1)
Pull Request Labels

Packages

  • Total packages: 1
  • Total downloads:
    • pypi 244 last-month
  • Total dependent packages: 1
  • Total dependent repositories: 0
  • Total versions: 45
  • Total maintainers: 1
pypi.org: tasksource

Preprocessings to prepare datasets for a task

  • Versions: 45
  • Dependent Packages: 1
  • Dependent Repositories: 0
  • Downloads: 244 Last month
Rankings
Dependent packages count: 6.6%
Downloads: 7.3%
Stargazers count: 12.1%
Average: 16.0%
Forks count: 23.2%
Dependent repos count: 30.6%
Maintainers (1)
sileod
Last synced: 6 months ago

Dependencies

.github/workflows/python-publish.yml actions
  • actions/checkout v3 composite
  • pypa/gh-action-pypi-publish release/v1 composite
.github/workflows/release.yml actions
  • actions/checkout v3 composite
pyproject.toml pypi