it return same vectors of "苹果" and "钢笔"

[damon-x]

I use nomic-embed-text:v1.5 in ollama ,I found that using the nomic-embed-text:v1.5 vector model, the word embeddings for '苹果' and '钢笔' turned out to be exactly the same.
'苹果' means apple in chinese
'钢笔' means pen in chinese
Embeddings for 'apple' and 'pen' are not same

the code is

def getEmbed(data):
    url = "http://127.0.0.1:11434/api/embed"
    rdata = {
       "model": "nomic-embed-text:v1.5",
       "input": data
    }
    payload = json.dumps(rdata)
    headers = {
       'Content-Type': 'application/json'
    }
    response = requests.request("POST", url, headers=headers, data=payload)
    resObj = json.loads(response.text)
    # print(resObj)
    return resObj['embeddings']


vecs = getEmbed(["苹果","钢笔"])
vec1 = vecs[0]
vec2 = vecs[1]
for i in range(0,len(vec1)):
    print(vec1[i] , vec2[i])