Final project for Ironhack Bootcamp of Data Analytics: NLP text generation for haikus and lyrics
This project contains three different NLP models with different approaches:
-
LSTM for Haikus generation Dataset: https://www.kaggle.com/hjhalani30/haiku-dataset This is the simpler one. A sequential model containing the following layers: Embedding Bidirectional LSTM Dense
-
LSTM for Lyrics generation Dataset: included in the repository Sequential model containing these layers: Embedding Bidirectional LSTM Dropout Unidirectional LSTM Embedding
-
GPT-2 for Haikus generation
Requeriments, prerequisites, dependencies, installation instructions.
└── final_project
├── .gitignore
├── README.md
├── LSTM_haiku
│ ├── raw_data
│ ├── processed_data
│ ├── LSTM_haiku_preprocessing.ipynb
│ └── LSTM_haiku_train.ipynb
│
├── LSTM_extremoduro
│ ├── raw_data
│ ├── processed_data
│ ├── LSTM_extremoduro_preprocessing.ipynb
│ └── LSTM_extremoduro_train.ipynb
│
└── gpt2_haiku
├── raw_data
├── processed_data
├── gpt2_haiku_preprocessing.ipynb
└── gpt2_haiku_train-ipynb
### **Update**
Added a Markov Chains model to generate song lyrics.
