DeepSeek Kotlin SDK/org.oremif.deepseek.models/ChatCompletionParams

ChatCompletionParams

class ChatCompletionParams : DeepSeekParams

Parameters for configuring chat completion requests to DeepSeek models.

This class encapsulates all the options available when sending chat completion requests, allowing fine-grained control over the model's behavior and output.

Example:

val params = chatCompletionParams {
    model = ChatModel.DEEPSEEK_CHAT
    temperature = 0.7
    maxTokens = 1000
    presencePenalty = 0.5
}

client.chat(params, messages)

Types

Builder

class Builder

Builder for creating ChatCompletionParams with standard (non-streaming) configuration.

StreamBuilder

class StreamBuilder

Builder for creating ChatCompletionParams specifically configured for streaming responses.

Properties

frequencyPenalty

val frequencyPenalty: Double? = null

Reduces repetition by penalizing tokens based on their frequency in the text

logprobs

val logprobs: Boolean? = null

Whether to return log probabilities of output tokens

maxTokens

val maxTokens: Int? = null

Maximum number of tokens to generate in the response

model

val model: ChatModel

The DeepSeek model to use for chat completion

presencePenalty

val presencePenalty: Double? = null

Reduces repetition by penalizing tokens that have appeared in the text

responseFormat

val responseFormat: ResponseFormat? = null

Format specification for the model's output

stop

val stop: StopReason? = null

Custom stop sequence that causes the model to stop generating further tokens

stream

val stream: Boolean? = null

Whether to stream the response back piece by piece

streamOptions

val streamOptions: StreamOptions? = null

Configuration options for streaming responses

temperature

val temperature: Double? = null

Controls randomness in output generation (higher = more random)

toolChoice

val toolChoice: ToolChoice? = null

Controls how the model selects tools to use

tools

val tools: List<Tool>? = null

List of tools that the model may use during chat completion

topLogprobs

val topLogprobs: Int? = null

How many most likely tokens to return at each position (max 20)

topP

val topP: Double? = null

Controls diversity via nucleus sampling (consider only tokens with top_p probability mass)

Functions

chat

fun chat(block: ChatCompletionParams.Builder.() -> Unit): ChatCompletionParams

Creates chat completion parameters with custom settings.

chatStream

fun chatStream(block: ChatCompletionParams.StreamBuilder.() -> Unit): ChatCompletionParams

Creates chat completion stream parameters with custom settings.

copy

fun copy(model: ChatModel = this.model, frequencyPenalty: Double? = this.frequencyPenalty, maxTokens: Int? = this.maxTokens, presencePenalty: Double? = this.presencePenalty, responseFormat: ResponseFormat? = this.responseFormat, stop: StopReason? = this.stop, stream: Boolean? = this.stream, streamOptions: StreamOptions? = this.streamOptions, temperature: Double? = this.temperature, topP: Double? = this.topP, tools: List<Tool>? = this.tools, toolChoice: ToolChoice? = this.toolChoice, logprobs: Boolean? = this.logprobs, topLogprobs: Int? = this.topLogprobs): ChatCompletionParams

Creates a copy of these parameters with optional changes to specific properties.

createRequest

fun createRequest(messages: List<ChatMessage>): ChatCompletionRequest

Creates a ChatCompletionRequest from these parameters and the provided messages.

equals

open operator override fun equals(other: Any?): Boolean

fim

fun fim(block: FIMCompletionParams.Builder.() -> Unit): FIMCompletionParams

Creates file-in-the-middle (FIM) completion parameters with custom settings.

fimStream

fun fimStream(block: FIMCompletionParams.StreamBuilder.() -> Unit): FIMCompletionParams

Creates file-in-the-middle (FIM) completion stream parameters with custom settings.

hashCode

open override fun hashCode(): Int

toString

open override fun toString(): String