Task1: Word2Vec
About
This implementation basicly refereced tensorflow's word2vec implementation of the Skip-Gram model.And implement a Contninus Bag of Words(CBOW) model. According to the main difference of above two model, I just change the generate_batch(...) function in word2vec_fns.py.
Usage
1.Get embeddings,If this completes without error, you should see a file called CBOW_Embeddings.npy in the current directory:
> python3 word2vec_cbow.py
2.Plot embeddings:
> python3 plot_embeddings.py
And you should see this:
Task2: Chinese Word Segmenter
About
This is a chinese segmenter implemented using the Max-Probability model.
Usage
1.Get word dict
python CreateDict.py
- run segmenter
python Segmenter.py
To-do
Using some more complex model, sach as RCF, LSTM...etc.
