Hate speech detection

Step 1: pre-processing
- You need to preprocess the train and test data given in the github repo.
- make sure at the end you get text and label in the data
- after this combined the data (train + test). This is final dataset which will be used for experimentation

Step 2: Preparing data for experimentation
- divide the data into 5 folds having random_state = 1
- these five folds will go into models

Step 3: Training Baselines
- Train 3 machine learning model (Naive Bayes, Logistic Regression, SVM) and evaluate on 5 folds
- Train 3 deep learning model (DNN, CNN, LSTM) and can use any word embeddings which you feel would work.
- Train 3 BERT models for text classification (BERT, RoBERTa, DistilRoberta), for this you can use "simple transformers" library

Step 4: Reporting Results
- Report the average over 5 folds
- Metrics that needs to be reported
Macro - Precision, Recall, F1-score
Weighted - Precision, Recall, F1-score
Average confusion matrix
Loss and accuracy curves for deep learning models

Note: pre-processing and machine learning models' training has been completed. Only need to implement DL models and Bert Models.