title | emoji | colorFrom | colorTo | sdk | sdk_version | app_file | pinned |
---|---|---|---|---|---|---|---|
DataMeasurementsTool |
🤗 |
indigo |
red |
streamlit |
1.0.0 |
app.py |
false |
🚧 Doing Construction 🚧
For more information, check out out blog post!
After cloning (and potentially setting up your virtual environment), run:
pip install -r requirements.txt
This installs all the requirements for the tool.
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"
streamlit run app.py