Testing with different dimensionality, LR, and tokenizations for word / token embedding

[max-andreasen]

• Do tests with the LSTM model using different sets of trained embeddings.
• Can change the LR in the word2vec model.
• Can change the dimensionality of the embedding vectors.
• Can change the vocabulary size of the BPE tokenization to create longer or shorter token lengths.