This is the final project for ift6289 at UdeM.

We use a single GPU (NVIDIA GeForce 2080 Ti) to develop this system with:

• Anaconda 3 (python 3.7)
• torch 1.10.2+cu113
• sklearn
• nltk 3.7
• pypinyin 0.46.0
• jieba 0.42.1

(1) Using mixed poetry data and music data to train, and testing on music data.

python main_poetry.py train character
python main_poetry.py test character

(2) Using mixed tones data and music data to train, and testing on music data.

python main_poetry.py train tone
python main_poetry.py test tone

(2) Using mixed English jokes data and music data to train, and testing on music data.

python main_joke.py train joke
python main_joke.py test joke

(4) Only using language data to train.

python main_baseline.py train_test 

The original experimental results of our project are in the experiment_data.txt.