Open mikelane opened 6 years ago
Pick a genre and a decade to train an LSTM. If there just isn't enough data, use the entire corpus.
Perhaps instead of this, use transfer learning as per Slack...
I'll use the 100 years dataset. Try Rap Genre first, if that isn't successful-I'll try the other idea as discussed.
Pick a genre and a decade to train an LSTM. If there just isn't enough data, use the entire corpus.
Perhaps instead of this, use transfer learning as per Slack...