https://github.com/huggingface/data-measurements-tool

Developing tools to automatically analyze datasets
https://github.com/huggingface/data-measurements-tool

Last synced: about 1 month ago
JSON representation

Developing tools to automatically analyze datasets

• Host: GitHub
• URL: https://github.com/huggingface/data-measurements-tool
• Owner: huggingface
• License: apache-2.0
• Created: 2021-07-20T19:08:09.000Z (about 3 years ago)
• Default Branch: main
• Last Pushed: 2023-10-21T19:03:30.000Z (11 months ago)
• Last Synced: 2024-05-09T12:51:53.827Z (4 months ago)
• Language: Python
• Size: 55.2 MB
• Stars: 68
• Watchers: 40
• Forks: 14
• Open Issues: 11
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
---

title: DataMeasurementsTool

emoji: 🤗

colorFrom: indigo

colorTo: red

sdk: streamlit

sdk_version: 1.0.0

app_file: app.py

pinned: false

---

# Data Measurements Tool

🚧 Doing Construction 🚧

[![Generic badge](https://img.shields.io/badge/🤗-Open%20In%20Spaces-blue.svg)](https://huggingface.co/spaces/huggingface/data-measurements-tool)

For more information, check out out [blog post](https://huggingface.co/blog/data-measurements-tool)!

# How to run:

After cloning (and potentially setting up your virtual environment), run:

`pip install -r requirements.txt`

This installs all the requirements for the tool.

## Command Line Interface

From there, you can measure different aspects of different datasets by running `run_data_measurements.py` with different options.

The options specify the HF Dataset, the Dataset config, the Dataset columns being measured, the measurements to use, and further details about caching and saving.

To see the full list of options, do:

`python3 run_data_measurements.py -h` or `python3 run_data_measurements.py --help`

Example for hate_speech18 dataset:

         `python3 run_data_measurements.py --dataset="hate_speech18" --config="default" --split="train" --feature="text"`

Example for getting *just* the nPMI measurement from hate_speech18:

         `python3 run_data_measurements.py --dataset=hate_speech18 --config default --split train --feature text --calculation npmi`

Example for IMDB dataset:

         `python3 run_data_measurements.py --dataset="imdb" --config="plain_text" --split="train" --label_field="label" --feature="text"`

## User Interface

`streamlit run app.py`