The new Global Similarity Graph Television News Sentence Embeddings dataset has reached 190 million 512-dimension embedding vectors!

Learn More.