In this project, using a model based on Bert transformer, SMS will be detected as spam.
Use python3.7 and requirement.txt to implement the project.
We trained the model on two datasets whose file addresses are below
The first model is trained on the "sms.csv" dataset The second model is trained on the "ExAIS.csv" dataset Also, the first model has recorded an accuracy of 99.61 And the second model recorded an accuracy of 87.89
run this command
python main.py -i /home/example.csv -o /home/example_dir -e 15Options
- -i Input CSV file absolute path
- -o director for save weights
- -b batch size
- -e number of epoch for train
- -lr learning rate