openai-core/com.aallam.openai.api.completion/CompletionRequest

CompletionRequest

@Serializable

class CompletionRequest(val model: ModelId, val prompt: String? = null, val maxTokens: Int? = null, val temperature: Double? = null, val topP: Double? = null, val n: Int? = null, val logprobs: Int? = null, val echo: Boolean? = null, val stop: List<String>? = null, val presencePenalty: Double? = null, val frequencyPenalty: Double? = null, val bestOf: Int? = null, val logitBias: Map<String, Int>? = null, val user: String? = null, val suffix: String? = null)

A request for OpenAI to generate a predicted completion for a prompt. All fields are Optional.

documentation

Constructors

CompletionRequest

constructor(model: ModelId, prompt: String? = null, maxTokens: Int? = null, temperature: Double? = null, topP: Double? = null, n: Int? = null, logprobs: Int? = null, echo: Boolean? = null, stop: List<String>? = null, presencePenalty: Double? = null, frequencyPenalty: Double? = null, bestOf: Int? = null, logitBias: Map<String, Int>? = null, user: String? = null, suffix: String? = null)

Properties

bestOf

@SerialName(value = "best_of")

val bestOf: Int? = null

Generates bestOf completions server-side and returns the "best" (the one with the lowest log probability per token). Results cannot be streamed.

echo

@SerialName(value = "echo")

val echo: Boolean? = null

Echo back the prompt in addition to the completion.

frequencyPenalty

@SerialName(value = "frequency_penalty")

val frequencyPenalty: Double? = null

Number between 0 and 1 (default 0) that penalizes new tokens based on their existing frequency in the text so far. Decreases the model's likelihood to repeat the same line verbatim.

logitBias

@SerialName(value = "logit_bias")

val logitBias: Map<String, Int>? = null

Modify the likelihood of specified tokens appearing in the completion.

logprobs

@SerialName(value = "logprobs")

val logprobs: Int? = null

Include the log probabilities on the logprobs most likely tokens, as well the chosen tokens. For example, if logprobs is 10, the API will return a list of the 10 most likely tokens. The API will always return the logprob of the sampled token, so there may be up to logprobs+1 elements in the response.

maxTokens

@SerialName(value = "max_tokens")

val maxTokens: Int? = null

The maximum number of tokens to generate. Requests can use up to 2048 tokens shared between prompt and completion. (One token is roughly 4 characters for normal English text)

model

@SerialName(value = "model")

val model: ModelId

ID of the model to use.

@SerialName(value = "n")

val n: Int? = null

How many completions to generate for each prompt.

presencePenalty

@SerialName(value = "presence_penalty")

val presencePenalty: Double? = null

Number between 0 and 1 (default 0) that penalizes new tokens based on whether they appear in the text so far. Increases the model's likelihood to talk about new topics.

prompt

@SerialName(value = "prompt")

val prompt: String? = null

The prompt(s) to generate completions for, encoded as a string, a list of strings, or a list of token lists.

stop

@SerialName(value = "stop")

val stop: List<String>? = null

Up to 4 sequences where the API will stop generating further tokens. The returned text will not contain the stop sequence.

suffix

@SerialName(value = "suffix")

val suffix: String? = null

The suffix that comes after a completion of inserted text.

temperature

@SerialName(value = "temperature")

val temperature: Double? = null

What sampling temperature to use. Higher values means the model will take more risks. Try 0.9 for more creative applications, and 0 (argmax sampling) for ones with a well-defined answer.

topP

@SerialName(value = "top_p")

val topP: Double? = null

An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered.

user

@SerialName(value = "user")

val user: String? = null

A unique identifier representing your end-user, which will help OpenAI to monitor and detect abuse.