This is a small project i made with the aim to build a Model for Sentiment analysis for Tweets. The model must receive a tweet as input and be able to predict whether the tweet has a positive, negative or neutral impression.
The model is builded using [python lightning(https://www.pytorchlightning.ai/) which is an ultimate PyTorch research framework that scale models without the boilerplate.
For the training of our model, we use the following dataset foung on Kaggle.
Beside our model, we will also build a small web server to deploy our application through REST API. We used Flask to build our server and deploy it as a docker container.
- write more details about the model
- explain the step for data clean
- provide more hyperparameters for reproductibility.