Transfer learning and Pretrained models Project

In this Project, there is a notebook which operate in google colab lab because it use gpu/cpu. So, It is Natural Language Programming Project in which python and other data science skills are use.
I have to prepare the data and tokenize it into subwords, and finally use it as input to some pretrained models, for example BERT.
I have to:
implement BPE algorithm
use ELMO and compare it with word2vec embeddings
explore the usage of BERT
train a classifier using BERT embeddings to solve COLA classification task
use prepared pipelines