This repository contains my implementations of Language Model for Text Generation which are extended from My CS224d Assignment Solutions
Simply run
python generate_text.py
give it some starter text and enjoy text generaions from the system :-)
- 1 Layer LSTM, word_embedding_size = 50, hidden_state_size = 100 , Test perplexity on PTB: 147.298156738
Some examples as produced by the system:
STARTER TEXT> microsoft shares
GENERATED TEXT> microsoft shares produced N million francs in three two months of these small investors.
STARTER TEXT> ozone is
GENERATED TEXT> ozone is still changed from the space.
STARTER TEXT> microsoft
GENERATED TEXT> microsoft said taiwan mci communications corp. closed at $ N a share .
STARTER TEXT> share
GENERATED TEXT> share of aetna may have backed up in the stock market <unk>.