<p>您可以使用<code>np.argsort(...)</code>进行排序</p>
<pre><code>import tensorflow_hub as hub
from sklearn.metrics.pairwise import cosine_similarity
embed = hub.load("https://tfhub.dev/google/universal-sentence-encoder/4")
seq1 = ["I'd like an apple juice",
"An apple a day keeps the doctor away",
"Eat apple every day",
"We buy apples every week",
"We use machine learning for text classification",
"Text classification is subfield of machine learning"]
embeddings1 = embed(seq1)
seq2 = ["I'd like an orange juice",
"An orange a day keeps the doctor away",
"Eat orange every day",
"We buy orange every week",
"We use machine learning for document classification",
"Text classification is some subfield of machine learning"]
embeddings2 = embed(seq2)
a = cosine_similarity(embeddings1, embeddings2)
</code></pre>
<hr/>
<pre><code>def get_pairs(a, b):
a = np.array(a)
b = np.array(b)
c = np.array(np.meshgrid(a, b))
c = c.T.reshape(len(a), -1, 2)
return c
</code></pre>
<hr/>
<pre><code>pairs = get_pairs(seq1, seq2)
sorted_idx = np.argsort(a, axis=0)[..., None]
sorted_pairs = pairs[sorted_idx]
print(pairs[0, 0])
print(pairs[0, 1])
print(pairs[0, 2])
</code></pre>
<hr/>
<pre><code>["I'd like an apple juice" "I'd like an orange juice"]
["I'd like an apple juice" 'An orange a day keeps the doctor away']
["I'd like an apple juice" 'Eat orange every day']
</code></pre>