#稀疏向量

Starting from the question of what sparse and dense really mean, this works through the embedding pipeline, subwords and vocabularies, input length, and their practical uses.