This repository houses the code I wrote for detecting the ease of readability of a given text excerpt which is the main task of the currently running CommonLit Readability Prize Competition
The main task in this competition is to take a text excerpt and predict it's ease of readability score (which is a continuous value).
I have also made a full training notebook in this competition.
The main data consists of 2 important features (or columns): excerpt
and target
.
The excerpt
is a basically a text sentence with a corresponding target
value which denotes how "easy it is to read a sentence"
The data consist of 3 files: train.csv
, test.csv
and sample_submission.csv
.
If you want to train the model on this data as-is, then you would typically have to perform 2 steps:
First, download the data from here.
Now, take the downloaded .zip
file and extract it into a new folder: input/
.
Make sure the input/
folder is at the same directory level as the train.py
file.
To run the code in this repository, you need a lot of frameworks installed on your system.
Make sure you have enough space on your disk and Internet quota before you proceed.
$ pip install -r requirements.txt
If you have done the above steps right, then just running the train.py
script should not produce any errors.
To run training, open the terminal and change your working directory to the same level as the train.py
file.
Now, for training do:
$ python train.py
This should start training in a few seconds and you should see a progress bar.
If you are having problems related to anything, please open an Issue and I will be happy to help!
I hope you found my work useful! If you did, then please โญ this repository!