tmteam / word2vec.tools Goto Github PK

Hi,

I tried substituting the vector operations in Word2vec.Tools with those in MathNet.Numerics, and got some weird results. It seems the "distances" used by Word2vec.Tools should be called "similarities".

Cosine similarity:

similarity = 1 − distance

for similarities: 1 = similar, 0 = not similar

for distances: 0 = close, 1 = distant

Hope that makes sense.

For reference:

• Math.NET Numerics Cosine Distance function

I am trying to add multiple representation of words by looping through each token at a time but the value of representation is not changing.
`

public void GetSimilarWordsMultipleTokensQuery(List targetWords, string trained_vector_file_path, int similar_word_count)
{
var vocabulary = new Word2vec.Tools.Word2VecBinaryReader().Read(trained_vector_file_path);
var additionVocab = vocabulary[targetWords[0]];

        for (int i = 1; i < targetWords.Count; i++)
        {
            additionVocab.Add(vocabulary[targetWords[i]]);
        }

        var closestAddition = vocabulary.Distance(additionVocab, similar_word_count);
        Console.WriteLine("Top " + similar_word_count + " that are closest to word " + "target words" + " are:-");

        foreach (var neighbourWord in closestAddition)
        {
            Console.WriteLine(neighbourWord.Representation.WordOrNull +"\t\t"+neighbourWord.DistanceValue);
        }
    }`

The Add function is not adding the representation once it's assigned.
Can you please resolve this issue. Thanks.

The tool returns a list of vocabulary words which are similar to the given word. However, I want to measure the similarity between two given words. How do I achieve that?

can we generate vectors.bin or txt with this tools ?

hi

i have tried some input data formats for trying get word vectors, but i can not confirm the data i using was correct.
Please helping me to know what data format i should use.
thank a lot!

the data format i use as following:
9 20
we need to test for text word2vec
word2vec is great
we need to test for text word2vec

Hello,

Thanks for sharing the great tool to enable using word2vector from c#. I have a word2vector model in binary trained by Gensim in Python. Can this tool load the model?

Thanks,
Juhua