torcharrow.functional.get_cosine_similarity

torcharrow.functional.get_cosine_similarity(input_ids: ListColumn, input_id_scores: ListColumn, matching_ids: ListColumn, matching_id_scores: ListColumn)

返回由 input_ids 定义的向量（由 input_id_scores 加权）与由 matching_ids 定义的向量（由 matching_id_scores 加权）之间的余弦相似度。

参数：

• input_ids (第一个 ID 列表) –

• input_ids_scores (input_ids 的分数（权重）) –

• matching_ids (第二个 ID 列表) –

• matching_ids_scores (matching_ids 的分数（权重）) –

示例

>>> import torcharrow as ta
>>> from torcharrow import functional
>>> input_ids = ta.column([[1, 2, 3]])
>>> input_id_scores = ta.column([[1.0,2.0,3.0]])
>>> matching_ids = ta.column([[1,2,3]])
>>> matching_id_scores = ta.column([[5.0,4.0,3.0]])
>>> functional.get_cosine_similarity(input_ids, input_id_scores, matching_ids, matching_id_scores)
0  0.831522
dtype: Float32(nullable=True), length: 1, null_count: 0