Nhảy tới nội dung

Class: ReplicateLLM

Replicate LLM implementation used

Extends​

  • BaseLLM

Constructors​

new ReplicateLLM()​

new ReplicateLLM(init?): ReplicateLLM

Parameters​

• init?: Partial <ReplicateLLM> & object

Returns​

ReplicateLLM

Overrides​

BaseLLM.constructor

Source​

packages/core/src/llm/replicate_ai.ts:115

Properties​

chatStrategy​

chatStrategy: ReplicateChatStrategy

Source​

packages/core/src/llm/replicate_ai.ts:109

maxTokens?​

optional maxTokens: number

Source​

packages/core/src/llm/replicate_ai.ts:112

model​

model: "Llama-2-70b-chat-old" | "Llama-2-70b-chat-4bit" | "Llama-2-13b-chat-old" | "Llama-2-13b-chat-4bit" | "Llama-2-7b-chat-old" | "Llama-2-7b-chat-4bit" | "llama-3-70b-instruct" | "llama-3-8b-instruct"

Source​

packages/core/src/llm/replicate_ai.ts:108

replicateSession​

replicateSession: ReplicateSession

Source​

packages/core/src/llm/replicate_ai.ts:113

temperature​

temperature: number

Source​

packages/core/src/llm/replicate_ai.ts:110

topP​

topP: number

Source​

packages/core/src/llm/replicate_ai.ts:111

Accessors​

metadata​

get metadata(): object

Returns​

object

contextWindow​

contextWindow: number

maxTokens​

maxTokens: undefined | number

model​

model: "Llama-2-70b-chat-old" | "Llama-2-70b-chat-4bit" | "Llama-2-13b-chat-old" | "Llama-2-13b-chat-4bit" | "Llama-2-7b-chat-old" | "Llama-2-7b-chat-4bit" | "llama-3-70b-instruct" | "llama-3-8b-instruct"

temperature​

temperature: number

tokenizer​

tokenizer: undefined = undefined

topP​

topP: number

Source​

packages/core/src/llm/replicate_ai.ts:140

Methods​

chat()​

chat(params)​

chat(params): Promise<AsyncIterable <ChatResponseChunk>>

Parameters​

• params: LLMChatParamsStreaming<object, object>

Returns​

Promise<AsyncIterable <ChatResponseChunk>>

Overrides​

BaseLLM.chat

Source​

packages/core/src/llm/replicate_ai.ts:307

chat(params)​

chat(params): Promise <ChatResponse<object>>

Parameters​

• params: LLMChatParamsNonStreaming<object, object>

Returns​

Promise <ChatResponse<object>>

Overrides​

BaseLLM.chat

Source​

packages/core/src/llm/replicate_ai.ts:310

complete()​

complete(params)​

complete(params): Promise<AsyncIterable <CompletionResponse>>

Parameters​

• params: LLMCompletionParamsStreaming

Returns​

Promise<AsyncIterable <CompletionResponse>>

Inherited from​

BaseLLM.complete

Source​

packages/core/src/llm/base.ts:22

complete(params)​

complete(params): Promise <CompletionResponse>

Parameters​

• params: LLMCompletionParamsNonStreaming

Returns​

Promise <CompletionResponse>

Inherited from​

BaseLLM.complete

Source​

packages/core/src/llm/base.ts:25

mapMessageTypeA16Z()​

mapMessageTypeA16Z(messageType): string

Parameters​

• messageType: MessageType

Returns​

string

Source​

packages/core/src/llm/replicate_ai.ts:217

mapMessagesToPrompt()​

mapMessagesToPrompt(messages): object

Parameters​

• messages: ChatMessage[]

Returns​

object

prompt​

prompt: string

systemPrompt​

systemPrompt: undefined | MessageContent

Source​

packages/core/src/llm/replicate_ai.ts:151

mapMessagesToPromptA16Z()​

mapMessagesToPromptA16Z(messages): object

Parameters​

• messages: ChatMessage[]

Returns​

object

prompt​

prompt: string

systemPrompt​

systemPrompt: undefined = undefined

Source​

packages/core/src/llm/replicate_ai.ts:203

mapMessagesToPromptLlama3()​

mapMessagesToPromptLlama3(messages): object

Parameters​

• messages: ChatMessage[]

Returns​

object

prompt​

prompt: string

systemPrompt​

systemPrompt: undefined = undefined

Source​

packages/core/src/llm/replicate_ai.ts:177

mapMessagesToPromptMeta()​

mapMessagesToPromptMeta(messages, opts?): object

Parameters​

• messages: ChatMessage[]

• opts?

• opts.replicate4Bit?: boolean

• opts.withBos?: boolean

• opts.withNewlines?: boolean

Returns​

object

prompt​

prompt: string

systemPrompt​

systemPrompt: undefined | MessageContent

Source​

packages/core/src/llm/replicate_ai.ts:230