Multilabel Text Classification with:
- TF-IDF + LogisiticRegression (baseline)
- Pretrained BertModel
The dataset is available at kaggle. It contains more than 310k Wikipedia comments which have been labeled by human raters for toxic behavior. The types of toxicity are:
- toxic
- severe_toxic
- obscene
- threat
- insult
- identity_hate
The user interface consists of file:
config.yaml- general configuration with data and model parameters
Default config.yaml:
seed: 42
data:
train_data_path: ../data/train.csv
test_data_path: ../data/test.csv
test_size: 0.05
sep: ','
text_column: comment_text
target_columns: ['toxic', 'severe_toxic', 'obscene', 'threat', 'insult', 'identity_hate']
output_path: output
log_file_path: logs.txt
model_type: baseline # bert (baseline = tf_idf + logreg)
save_model: true
test_run: true
tf-idf:
word:
sublinear_tf: true
strip_accents: unicode
analyzer: word
ngram_range: (1, 1)
max_features: 10000
char:
sublinear_tf: true
strip_accents: unicode
analyzer: char
ngram_range: (2, 6)
max_features: 50000
logreg:
penalty: l2
C: 1.0
class_weight: balanced
solver: lbfgs
n_jobs: -1
bert:
bert_model_name: bert-base-cased
max_token_len: 128
batch_size: 32
epochs: 2
learning_rate: 0.00002
warmup_frac: 0.2
Run in terminal: python -m Toxic_comment_classification
After training the model, the pipeline will return the following files:
logreg_model.joblib/bert_best_model.pt-- saved modelslogs.txt-- file with logs
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent