3 ms·
Out of curiosity related to the word vectorization algorithm...why does one word not perform as well? Whats the cause/rationale?
by Solvency 3y ago
Out of curiosity related to the word vectorization algorithm...why does one word not perform as well? Whats the cause/rationale?
- julien040 3y agoIt's pure speculation, but articles embeddings are computed using 512 tokens, which is roughly equivalent to 400 words. I think that using only one word does not allow the model to fully understand the context.