Completions

Article Summary

Share feedback

Thanks for sharing your feedback!

Available in Classic and VPC

You can pass information for generating a result value (completion) in CLOVA Studio to CLOVA Studio. CLOVA Studio creates a result value based on the specified contents of the prompt, the model to be used, and other parameter values.

Requests

Describes the request format. The request format is as follows:

Method	URI
POST	/v1/completions/{model}

Header

The following describes the header.

Header	Required	Description
X-NCP-CLOVASTUDIO-API-KEY	Y	API key issued when creating a test app or service app
X-NCP-APIGW-API-KEY	Y	API Gateway key issued when creating a test app or service app
X-NCP-CLOVASTUDIO-REQUEST-ID	N	Request ID for each request
Content-Type	Y	application/json

Path parameters

The following describes the parameters.

Field	Type	Required	Description
model	string	Y	Model name: `LK-D2`

Body

The following describes the body.

Field	Type	Required	Description
includeAiFilters	boolean	N	Indicates the degree to which the generated text corresponds to each category (for example, profanity, abasement, discrimination, hate speech, sexual harassment, or obscenity) true (default value) false
includeTokens	boolean	N	Whether to include input_tokens and output_tokens in the response true: use false: do not use
maxTokens	int	N	Maximum number of tokens to be used when generating result value 0 ＜ maxTokens ≤ 2048 (default value: 100)
repeatPenalty	float	N	Penalty value to be applied to tokens that repeat the same expression 0.0 ＜ repeatPenalty ≤ 10.0 (default value: 5.0)
restart	string	N	Text to be appended after output value
start	string	N	Text to be appended after user input value
stopBefore	array	N	Character that halts token generation [] (default value)
temperature	float	N	Randomness of result values. A higher value is more likely to produce random results 0.00 ＜ temperature ≤ 1.00 (default value: 0.50)
text	string	Y	It is a user input value (prompt), and a new sentence is generated based on this text
topK	int	N	Top K value. Tokens are listed in order of highest probability, and only tokens within the top K range are used 0 ≤ topk ≤ 128 (default value: 0)
topP	float	N	Top P value. Select a token with top P included in its cumulative probability value as a candidate 0.00 ＜ topP ≤ 1.00 (default value: 0.80)

Syntax

The following is an example of syntax.

curl --location --request POST 'https://clovastudio.apigw.ntruss.com/testapp/v1/completions/LK-D2' \
--header 'X-NCP-CLOVASTUDIO-API-KEY: <X-NCP-CLOVASTUDIO-API-KEY>' \
--header 'X-NCP-APIGW-API-KEY: <X-NCP-APIGW-API-KEY>' \
--header 'X-NCP-CLOVASTUDIO-REQUEST-ID: <X-NCP-CLOVASTUDIO-REQUEST-ID>' \
--header 'Content-Type: application/json' \
--data '{
 "start": "\nEnglish: ",
  "stopBefore": ["\nKorean: "],
  "text": "Korean: apple\nEnglish: apple\nKorean: banana",
  "includeAiFilters": true
}'

Responses

Describes the responses format.

Header

The following describes the header.

Header	Required	Description
Content-Type	-	application/json

Body

The following describes the body.

Field	Type	Required	Description
aiFilters	object array	-	AI Filter score
aiFilters[].groupName	string	-	See AI Filter table
aiFilters[].name	string	-	See AI Filter table
aiFilters[].score	string	-	The extent to which the text corresponds to the aiFilters[].name category 0: high 1: medium 2: low
inputLength	integer	-	Input token number
inputTokens	string array	-	Input token information
outputLength	integer	-	Output token number
outputTokens	string array	-	Output token information
stopReason	string	-	Reason for stopping result value length: reached the set maximum length end_token: reached the maximum number of tokens that can be used stop_before: reached the set stop sequence
text	string	-	Generated token. Includes both prompt and result values
probs	float	-	Probability of each generated token being selected

AI Filter table

See the following table for contents included in aiFilters[].groupName and aiFilters[].name.

groupName	name	Description
curse	discrimination	Abasement, discrimination, or hate speech
curse	insult	Profanity
unsafeContents	sexualHarassment	Sexual harassment or obscenity

Syntax

The following is an example of syntax.

{
  "status": {
    "code": "20000",
    "message": "OK"
  },
  "result": {
    "text": "...",
    "stopReason": "stop_before",
    "inputLength": 10,
    "outputLength": 20,
    "inputTokens" : ["who", " ", "are", " ", "you"],
    "outputTokens" : ["hello", " ", "world"],
    "aiFilters": [
      {
        "groupName": "curse",
        "name": "insult",
        "score": "0"
      },
      {
        "groupName": "curse",
        "name": "discrimination",
        "score": "1"
      },
      {
        "groupName": "unsafeContents",
        "name": "sexualHarassment",
        "score": "2"
      }
    ]
  }
}

Was this article helpful?

What's Next

Generating training

Table of contents

Requests
Responses