Qwen Embedding v3 API nasıl kullanılır?

Qwen Embedding v3 modelini Parel API ile kullanmak için OpenAI SDK'sını yükleyin, base_url olarak https://api.parel.cloud/v1 belirleyin ve model parametresine "qwen-embedding-v3" yazın. Python, Node.js, Java, C# ve cURL desteklenir.

Qwen Embedding v3 KVKK uyumlu mu?

Qwen Embedding v3 DashScope (Global) sunucularında çalışır. KVKK uyumlu alternatifler için Parel edge modellerine bakabilirsiniz.

Alibaba

Qwen Embedding v3

Semantik arama ve RAG pipeline'ları için

Input 7 kredi/1M token

Output 0 kredi/1M token

Playground'da Dene

Embedding 8K token

Hakkında

Alibaba cok dilli embedding. 1024 boyut, 50+ dil. RAG icin ideal.

Kullanım Alanları

› Semantik arama ve belge erişimi
› RAG (Retrieval-Augmented Generation) pipeline'ları
› Metin benzerliği ve kümeleme analizi
› Öneri sistemleri ve kişiselleştirme
› Anlamsal metin sınıflandırması

Aile

Qwen Embedding

Mimari

Transformer

Veri Merkezi

DashScope (Global)

Provider: Alibaba
Tür: Embedding
Boyut: 8K token
Model ID: qwen-embedding-v3
Input Fiyat: 7 kredi/1M token
Output Fiyat: 0 kredi/1M token

Benzer Modeller

Amazon Titan Embed v2

Amazon · 8K

Daha ucuz

Gemini Embedding 2

Google · 8K

Daha pahalı, daha güçlü

Text Embedding 3 Small

OpenAI · 8K

Daha pahalı, daha güçlü

OpenAI SDK ile dakikalar içinde entegre edin.

Ücretsiz Başla Tüm Modeller

Authentication #

API anahtarınızı ortam değişkeni olarak tanımlayın.

python

import os
os.environ["OPENAI_API_KEY"] = "pk-..."

Kurulum #

OpenAI SDK'yi yükleyin.

python

pip install openai

Embedding Oluştur #

Tek bir metin için vektörel temsil oluşturun.

python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.parel.cloud/v1",
    api_key="pk-...",
)

response = client.embeddings.create(
    model="qwen-embedding-v3",
    input="Ornek metin",
)

embedding = response.data[0].embedding
print(f"Boyut: {len(embedding)}")
print(embedding[:5])

javascript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.parel.cloud/v1",
  apiKey: "pk-...",
});

const response = await client.embeddings.create({
  model: "qwen-embedding-v3",
  input: "Ornek metin",
});

const embedding = response.data[0].embedding;
console.log("Boyut:", embedding.length);
console.log(embedding.slice(0, 5));

java

import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.*;

OpenAIClient client = OpenAIOkHttpClient.builder()
    .baseUrl("https://api.parel.cloud/v1")
    .apiKey("pk-...")
    .build();

EmbeddingCreateParams params = EmbeddingCreateParams.builder()
    .model("qwen-embedding-v3")
    .input("Ornek metin")
    .build();

CreateEmbeddingResponse response = client.embeddings().create(params);

List<Double> embedding = response.data().get(0).embedding();
System.out.println("Boyut: " + embedding.size());
System.out.println(embedding.subList(0, 5));

csharp

using OpenAI;
using OpenAI.Embeddings;

EmbeddingClient client = new(
    model: "qwen-embedding-v3",
    credential: new ApiKeyCredential("pk-..."),
    options: new OpenAIClientOptions {
        Endpoint = new Uri("https://api.parel.cloud/v1") }
);

OpenAIEmbedding embedding = client.GenerateEmbedding("Ornek metin");

ReadOnlyMemory<float> vector = embedding.ToFloats();
Console.WriteLine($"Boyut: {vector.Length}");
Console.WriteLine(string.Join(", ", vector.Span.Slice(0, 5).ToArray()));

bash

curl -X POST https://api.parel.cloud/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer pk-..." \
  -d '{
    "model": "qwen-embedding-v3",
    "input": "Ornek metin"
  }'

Toplu Embedding #

Birden fazla metin icin tek istekte embedding alin.

python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.parel.cloud/v1",
    api_key="pk-...",
)

response = client.embeddings.create(
    model="qwen-embedding-v3",
    input=[
        "Birinci metin",
        "Ikinci metin",
        "Ucuncu metin",
    ],
)

for i, item in enumerate(response.data):
    print(f"[{i}] Boyut: {len(item.embedding)}")

javascript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.parel.cloud/v1",
  apiKey: "pk-...",
});

const response = await client.embeddings.create({
  model: "qwen-embedding-v3",
  input: [
    "Birinci metin",
    "Ikinci metin",
    "Ucuncu metin",
  ],
});

response.data.forEach((item, i) => {
  console.log(`[${i}] Boyut: ${item.embedding.length}`);
});

java

import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.*;
import java.util.List;

OpenAIClient client = OpenAIOkHttpClient.builder()
    .baseUrl("https://api.parel.cloud/v1")
    .apiKey("pk-...")
    .build();

EmbeddingCreateParams params = EmbeddingCreateParams.builder()
    .model("qwen-embedding-v3")
    .input(List.of("Birinci metin", "Ikinci metin", "Ucuncu metin"))
    .build();

CreateEmbeddingResponse response = client.embeddings().create(params);

for (int i = 0; i < response.data().size(); i++) {
    System.out.printf("[%d] Boyut: %d%n", i, response.data().get(i).embedding().size());
}

csharp

using OpenAI;
using OpenAI.Embeddings;

EmbeddingClient client = new(
    model: "qwen-embedding-v3",
    credential: new ApiKeyCredential("pk-..."),
    options: new OpenAIClientOptions {
        Endpoint = new Uri("https://api.parel.cloud/v1") }
);

OpenAIEmbeddingCollection result = client.GenerateEmbeddings(
    new List<string> { "Birinci metin", "Ikinci metin", "Ucuncu metin" }
);

for (int i = 0; i < result.Count; i++)
{
    Console.WriteLine($"[{i}] Boyut: {result[i].ToFloats().Length}");
}

bash

curl -X POST https://api.parel.cloud/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer pk-..." \
  -d '{
    "model": "qwen-embedding-v3",
    "input": ["Birinci metin", "Ikinci metin", "Ucuncu metin"]
  }'

Cosine Similarity #

İki metin arasındaki benzerlik skorunu hesaplayın.

python

import numpy as np
from openai import OpenAI

client = OpenAI(
    base_url="https://api.parel.cloud/v1",
    api_key="pk-...",
)

def get_embedding(text: str) -> list[float]:
    response = client.embeddings.create(
        model="qwen-embedding-v3",
        input=text,
    )
    return response.data[0].embedding

def cosine_similarity(a, b):
    a, b = np.array(a), np.array(b)
    return np.dot(a, b) / (np.linalg.norm(a) * np.linalg.norm(b))

e1 = get_embedding("Istanbul'un nufusu kac?")
e2 = get_embedding("Istanbul'da kac kisi yasiyor?")
e3 = get_embedding("Bugun hava nasil?")

print(f"Benzer: {cosine_similarity(e1, e2):.4f}")
print(f"Farkli: {cosine_similarity(e1, e3):.4f}")

javascript

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.parel.cloud/v1",
  apiKey: "pk-...",
});

async function getEmbedding(text) {
  const res = await client.embeddings.create({
    model: "qwen-embedding-v3",
    input: text,
  });
  return res.data[0].embedding;
}

function cosineSimilarity(a, b) {
  let dot = 0, normA = 0, normB = 0;
  for (let i = 0; i < a.length; i++) {
    dot += a[i] * b[i];
    normA += a[i] * a[i];
    normB += b[i] * b[i];
  }
  return dot / (Math.sqrt(normA) * Math.sqrt(normB));
}

const e1 = await getEmbedding("Istanbul'un nufusu kac?");
const e2 = await getEmbedding("Istanbul'da kac kisi yasiyor?");
const e3 = await getEmbedding("Bugun hava nasil?");

console.log("Benzer:", cosineSimilarity(e1, e2).toFixed(4));
console.log("Farkli:", cosineSimilarity(e1, e3).toFixed(4));

java

import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.models.*;

OpenAIClient client = OpenAIOkHttpClient.builder()
    .baseUrl("https://api.parel.cloud/v1")
    .apiKey("pk-...")
    .build();

List<Double> getEmbedding(String text) {
    EmbeddingCreateParams params = EmbeddingCreateParams.builder()
        .model("qwen-embedding-v3")
        .input(text)
        .build();
    return client.embeddings().create(params).data().get(0).embedding();
}

double cosineSimilarity(List<Double> a, List<Double> b) {
    double dot = 0, normA = 0, normB = 0;
    for (int i = 0; i < a.size(); i++) {
        dot += a.get(i) * b.get(i);
        normA += a.get(i) * a.get(i);
        normB += b.get(i) * b.get(i);
    }
    return dot / (Math.sqrt(normA) * Math.sqrt(normB));
}

var e1 = getEmbedding("Istanbul'un nufusu kac?");
var e2 = getEmbedding("Istanbul'da kac kisi yasiyor?");
var e3 = getEmbedding("Bugun hava nasil?");

System.out.printf("Benzer: %.4f%n", cosineSimilarity(e1, e2));
System.out.printf("Farkli: %.4f%n", cosineSimilarity(e1, e3));

csharp

using OpenAI;
using OpenAI.Embeddings;

EmbeddingClient client = new(
    model: "qwen-embedding-v3",
    credential: new ApiKeyCredential("pk-..."),
    options: new OpenAIClientOptions {
        Endpoint = new Uri("https://api.parel.cloud/v1") }
);

float[] GetEmbedding(string text) =>
    client.GenerateEmbedding(text).ToFloats().ToArray();

float CosineSimilarity(float[] a, float[] b)
{
    float dot = 0, normA = 0, normB = 0;
    for (int i = 0; i < a.Length; i++)
    {
        dot += a[i] * b[i];
        normA += a[i] * a[i];
        normB += b[i] * b[i];
    }
    return dot / (MathF.Sqrt(normA) * MathF.Sqrt(normB));
}

var e1 = GetEmbedding("Istanbul'un nufusu kac?");
var e2 = GetEmbedding("Istanbul'da kac kisi yasiyor?");
var e3 = GetEmbedding("Bugun hava nasil?");

Console.WriteLine($"Benzer: {CosineSimilarity(e1, e2):F4}");
Console.WriteLine($"Farkli: {CosineSimilarity(e1, e3):F4}");

bash

# 1. Embedding'leri alin
curl -X POST https://api.parel.cloud/v1/embeddings \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer pk-..." \
  -d '{
    "model": "qwen-embedding-v3",
    "input": ["Istanbul nufusu kac?", "Istanbul kac kisi?"]
  }'

# 2. Cosine similarity hesabini kendi kodunuzda yapin

Kayıt olun, API anahtarınızı alın.

Hemen Başla Tum Modeller