Static Models

Static semantic models

In this section, we will present the different static semantic models or word embedding representations we have built for Amharic

The Amharic word2Vec model is available here

To use the model, first download the model

wget http://ltdata1.informatik.uni-hamburg.de/amharic/models/word2vec/amharic-word2vec-300D.gz

Or use any other tool. Once the amharic-word2vec-300D.gz model is downloaded, run the following code

import gensim
model = gensim.models.KeyedVectors.load_word2vec_format('PATH_TO/amharic-word2vec-300D.gz', binary=False) 
model.most_similar('በሬ')

You should get results as follows

Amharic word2Vec

The Amharic fastText model is available here

To use the model, first download the model

wget http://ltdata1.informatik.uni-hamburg.de/amharic/models/fasttext/amharic-fasttext-300D.gz

Or use any other tool. Once the amharic-fasttext-300D.gz model is downloaded, run the following code

import gensim
model = gensim.models.KeyedVectors.load_word2vec_format('PATH_TO/amharic-fasttext-300D.gz', binary=False) 
model.most_similar('በሬ')