Myers Briggs Type Indicator personality Classification based on users' from social media posts.
To get a local copy up and running follow these simple steps.
- pip
pip install torch pandas transformers tokenizers datasets numpy
Here the user has to input the required command line arguments to run the model.
It is recommended for an intial user to run "python main.py --help" , for information on each argument.
"-m" -> Model name.
"-d" -> Dataset directory path.
"-dl" -> Next delimiter.
"-w" -> Minimum words per post.
"-tb" -> Training batch size.
"-eb" -> Eval batch size.
"-r" -> Test split ratio.
"-l" -> Learning rate.
"-wd" -> Weight decay.
"-n" -> Num classes.
"-ml"-> Multilabel classification.
"-f" -> Number of epochs.
"-mt"-> Metrics.
"--optimizer" -> Optimizer.
"--loss" -> Loss criterion.
"--train"->To train model.
"-o"-> Output directory .
"--save_interval" python main.py -m vinai/bertweet-base -d ./mbti_1.csv -dl "|||" -w 5 -tb 256 -eb 256 -r 0.75 -l 176e-06 -wd 1e-05 -n 4 -f 5 --optimizer ADAM --loss BCE --train -o ./ops --save_interval 2