> ## Documentation Index
> Fetch the complete documentation index at: https://apidoc.cometapi.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Tạo embeddings
> Sử dụng CometAPI POST /v1/embeddings để tạo vector embeddings cho dữ liệu đầu vào văn bản với một embedding model được chọn.
CometAPI hỗ trợ các embedding model từ nhiều nhà cung cấp thông qua một endpoint duy nhất. Truyền vào một hoặc nhiều chuỗi văn bản và nhận về các vector số cho semantic search, phân cụm, phân loại hoặc retrieval-augmented generation (RAG). Xem [danh sách model](/vi/overview/models) để biết các embedding model khả dụng và giá.
Các model `text-embedding-3-*` hỗ trợ tham số `dimensions`, cho phép rút gọn vector embedding mà không làm giảm đáng kể độ chính xác. Điều này có thể giúp giảm chi phí lưu trữ trong khi vẫn giữ lại phần lớn thông tin ngữ nghĩa.
Để nhúng nhiều văn bản trong một request, hãy truyền một mảng chuỗi vào tham số `input`. Batch input hiệu quả hơn đáng kể so với việc thực hiện từng request riêng lẻ.
***
## OpenAPI
````yaml api/openapi/text/post-embeddings.openapi.json POST /v1/embeddings
openapi: 3.1.0
info:
title: Embeddings API
version: 1.0.0
servers:
- url: https://api.cometapi.com
security:
- bearerAuth: []
paths:
/v1/embeddings:
post:
summary: Create Embeddings
operationId: createEmbeddings
requestBody:
required: true
content:
application/json:
schema:
type: object
required:
- model
- input
properties:
model:
type: string
description: >-
The embedding model to use. See the [Models
page](/overview/models) for current embedding model IDs.
example: text-embedding-3-small
input:
oneOf:
- type: string
description: A single text string to embed.
- type: array
description: >-
An array of strings to embed in a single request. Each
string can be up to 8,191 tokens.
items:
type: string
- type: array
description: An array of token arrays.
items:
type: array
items:
type: integer
description: >-
The text to embed. Can be a single string, an array of
strings, or an array of token arrays. Each input must not
exceed the model's maximum token limit (8,191 tokens for
`text-embedding-3-*` models).
encoding_format:
type: string
description: >-
The format of the returned embedding vectors. `float`
returns an array of floating-point numbers. `base64` returns
a base64-encoded string representation, which can reduce
response size for large batches.
enum:
- float
- base64
default: float
dimensions:
type: integer
description: >-
The number of dimensions for the output embedding vector.
Only supported by `text-embedding-3-*` models. Reducing
dimensions can lower storage costs while maintaining most of
the embedding's utility.
minimum: 1
user:
type: string
description: >-
A unique identifier for your end-user, which can help
monitor and detect abuse.
examples:
Single Text:
summary: Single Text
value:
model: text-embedding-3-small
input: The food was delicious and the waiter was friendly.
Batch Input:
summary: Batch Input
value:
model: text-embedding-3-small
input:
- Hello world
- How are you?
- Embedding example
encoding_format: float
With Dimensions:
summary: Reduced Dimensions
value:
model: text-embedding-3-small
input: Search query text
dimensions: 256
responses:
'200':
description: A list of embedding vectors for the input text(s).
content:
application/json:
schema:
type: object
properties:
object:
type: string
description: The object type, always `list`.
enum:
- list
example: list
data:
type: array
description: >-
An array of embedding objects, one per input text. When
multiple inputs are provided, results are returned in the
same order as the input.
items:
type: object
properties:
object:
type: string
description: The object type, always `embedding`.
enum:
- embedding
example: embedding
index:
type: integer
description: >-
The index of this embedding in the input array
(starting from 0).
example: 0
embedding:
type: array
description: >-
The embedding vector as an array of floating-point
numbers. The length depends on the model and
`dimensions` parameter.
items:
type: number
example:
- -0.0021
- -0.0491
- 0.0209
- 0.0314
- -0.0453
model:
type: string
description: The model used to generate the embeddings.
example: text-embedding-3-small
usage:
type: object
description: Token usage statistics for this request.
properties:
prompt_tokens:
type: integer
description: The number of tokens in the input text(s).
example: 2
total_tokens:
type: integer
description: >-
The total number of tokens processed (same as
`prompt_tokens` for embeddings).
example: 2
example:
object: list
data:
- object: embedding
index: 0
embedding:
- -0.0021
- -0.0491
- 0.0209
- 0.0314
- -0.0453
model: text-embedding-3-small
usage:
prompt_tokens: 2
total_tokens: 2
x-codeSamples:
- lang: Python
label: Single Text
source: |
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.cometapi.com/v1",
api_key=os.environ["COMETAPI_KEY"],
)
response = client.embeddings.create(
model="text-embedding-3-small",
input="The food was delicious and the waiter was friendly.",
)
print(response.data[0].embedding[:5]) # First 5 dimensions
print(f"Dimensions: {len(response.data[0].embedding)}")
- lang: Python
label: Batch Input
source: |
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.cometapi.com/v1",
api_key=os.environ["COMETAPI_KEY"],
)
response = client.embeddings.create(
model="text-embedding-3-small",
input=["Hello world", "How are you?", "Embedding example"],
)
for item in response.data:
print(f"Index {item.index}: {len(item.embedding)} dimensions")
- lang: Python
label: Reduced Dimensions
source: |
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.cometapi.com/v1",
api_key=os.environ["COMETAPI_KEY"],
)
# Use fewer dimensions to reduce storage costs
response = client.embeddings.create(
model="text-embedding-3-small",
input="Search query text",
dimensions=256,
)
print(f"Dimensions: {len(response.data[0].embedding)}") # 256
- lang: JavaScript
label: Single Text
source: |
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.cometapi.com/v1",
apiKey: process.env.COMETAPI_KEY,
});
const response = await client.embeddings.create({
model: "text-embedding-3-small",
input: "The food was delicious and the waiter was friendly.",
});
console.log(response.data[0].embedding.slice(0, 5));
console.log(`Dimensions: ${response.data[0].embedding.length}`);
- lang: JavaScript
label: Batch Input
source: |
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.cometapi.com/v1",
apiKey: process.env.COMETAPI_KEY,
});
const response = await client.embeddings.create({
model: "text-embedding-3-small",
input: ["Hello world", "How are you?", "Embedding example"],
});
for (const item of response.data) {
console.log(`Index ${item.index}: ${item.embedding.length} dimensions`);
}
- lang: JavaScript
label: Reduced Dimensions
source: >
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.cometapi.com/v1",
apiKey: process.env.COMETAPI_KEY,
});
// Use fewer dimensions to reduce storage costs
const response = await client.embeddings.create({
model: "text-embedding-3-small",
input: "Search query text",
dimensions: 256,
});
console.log(`Dimensions: ${response.data[0].embedding.length}`); //
256
- lang: Shell
label: Single Text
source: |
curl https://api.cometapi.com/v1/embeddings \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $COMETAPI_KEY" \
-d '{
"model": "text-embedding-3-small",
"input": "The food was delicious and the waiter was friendly."
}'
- lang: Shell
label: Batch Input
source: |
curl https://api.cometapi.com/v1/embeddings \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $COMETAPI_KEY" \
-d '{
"model": "text-embedding-3-small",
"input": ["Hello world", "How are you?", "Embedding example"]
}'
- lang: Shell
label: Reduced Dimensions
source: |
curl https://api.cometapi.com/v1/embeddings \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $COMETAPI_KEY" \
-d '{
"model": "text-embedding-3-small",
"input": "Search query text",
"dimensions": 256
}'
components:
securitySchemes:
bearerAuth:
type: http
scheme: bearer
description: Bearer token authentication. Use your CometAPI key.
````