metadata
language: en
tags:
- grammar
- spell
- correction
Spellcheck Model
Installation
- Install Python 3.8 or higher and Docker;
- Clone the repo;
- Create a virtual environment:
python3 -m venv venv; - Activate the virtual environment:
source venv/bin/activate; - Install dependencies:
pip install -r requirements.txt;
Usage
Training
There's a data directory with two files:
eval.csv: evaluation dataset;train.csv: training dataset; Both are in the same format:
input,target
"grammar: som sentense","some sentence"
"grammar: anoder centence","another sentence"
...
You can edit them to use your own dataset.
Now if you want to fine-tune the existing model here, just run the following commands:
docker build -t spellcheck_train .
docker run -v $(pwd):/app spellcheck_train
Also make sure you have docker running on your machine.
Testing
If you want to test your model, just run the following command:
python test.py