Skip to content

Commit

Permalink
[editor] Add AnyscaleEndpoint Prompt Schema
Browse files Browse the repository at this point in the history
  • Loading branch information
Ryan Holinshead committed Jan 5, 2024
1 parent f59b724 commit a9784ad
Show file tree
Hide file tree
Showing 3 changed files with 109 additions and 0 deletions.
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@
import { PromptSchema } from "../../utils/promptUtils";

export const AnyscaleEndpointPromptSchema: PromptSchema = {
// See https://docs.anyscale.com/endpoints/model-serving/openai-migration-guide#step-3-check-parameter-compatibility
// for settings and defaults. The settings below are supported settings specified in the OpenAIInference
// refine_chat_completion_params implementation.
input: {
type: "string",
},
model_settings: {
type: "object",
properties: {
model: {
type: "string",
},
frequency_penalty: {
type: "number",
minimum: -2.0,
maximum: 2.0,
description: `Number between -2.0 and 2.0.
Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.`,
},
max_tokens: {
type: "integer",
description: `The maximum number of tokens to generate in the chat completion.`,
},
presence_penalty: {
type: "number",
minimum: -2.0,
maximum: 2.0,
description: `Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far,
increasing the model's likelihood to talk about new topics.`,
},
stop: {
type: "array",
items: {
type: "string",
},
description: `Up to 4 sequences where the API will stop generating further tokens. The returned text will not contain the stop sequence.`,
},
stream: {
type: "boolean",
description: `If true, send messages token by token. If false, messages send in bulk.`,
},
temperature: {
type: "number",
minimum: 0.0,
maximum: 2.0,
description: `A number between 0 and 2. Higher values correspond to more random responses and lower values being more deterministic.`,
},
top_p: {
type: "number",
minimum: 0.0,
maximum: 1.0,
description: `The percentage of tokens with top_p probability mass to consider.
For example, 0.1 means only tokens comprising the top 10% probability mass become candidates.`,
},
},
},
prompt_metadata: {
type: "object",
properties: {
remember_chat_context: {
type: "boolean",
},
},
},
};
4 changes: 4 additions & 0 deletions python/src/aiconfig/editor/client/src/utils/promptUtils.ts
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ import { DalleImageGenerationParserPromptSchema } from "../shared/prompt_schemas
import { PaLMTextParserPromptSchema } from "../shared/prompt_schemas/PaLMTextParserPromptSchema";
import { PaLMChatParserPromptSchema } from "../shared/prompt_schemas/PaLMChatParserPromptSchema";
import { HuggingFaceTextGenerationParserPromptSchema } from "../shared/prompt_schemas/HuggingFaceTextGenerationParserPromptSchema";
import { AnyscaleEndpointPromptSchema } from "../shared/prompt_schemas/AnyscaleEndpointPromptSchema";

/**
* Get the name of the model for the specified prompt. The name will either be specified in the prompt's
Expand Down Expand Up @@ -76,6 +77,9 @@ export const PROMPT_SCHEMAS: Record<string, PromptSchema> = {

// PaLMChatParser
"models/chat-bison-001": PaLMChatParserPromptSchema,

// AnyscaleEndpoint
AnyscaleEndpoint: AnyscaleEndpointPromptSchema,
};

export type PromptInputSchema =
Expand Down
37 changes: 37 additions & 0 deletions python/src/aiconfig/editor/travel.aiconfig.json
Original file line number Diff line number Diff line change
Expand Up @@ -94,6 +94,43 @@
}
}
]
},
{
"name": "prompt_4",
"input": "test",
"metadata": {
"model": {
"name": "AnyscaleEndpoint",
"settings": {
"model": "meta-llama/Llama-2-7b-chat-hf"
}
},
"parameters": {}
},
"outputs": [
{
"output_type": "execute_result",
"execution_count": 0,
"data": " Sure! Here are 5 questions to test your understanding of the conversation:\n\n1. What is the main topic of the conversation?\n2. What is the purpose of the meeting according to the speaker?\n3. What is the speaker's opinion on the proposed meeting time?\n4. How does the speaker feel about the potential attendees of the meeting?\n5. What is the speaker's recommended solution for the meeting time conflict?",
"metadata": {
"raw_response": {
"content": " Sure! Here are 5 questions to test your understanding of the conversation:\n\n1. What is the main topic of the conversation?\n2. What is the purpose of the meeting according to the speaker?\n3. What is the speaker's opinion on the proposed meeting time?\n4. How does the speaker feel about the potential attendees of the meeting?\n5. What is the speaker's recommended solution for the meeting time conflict?",
"role": "assistant"
},
"id": "meta-llama/Llama-2-7b-chat-hf-JknFQzgthe2aABGHrYeW86GhXnUVx7fH9Z7JldFLe2g",
"created": 1704484147,
"model": "meta-llama/Llama-2-7b-chat-hf",
"object": "text_completion",
"usage": {
"completion_tokens": 97,
"prompt_tokens": 9,
"total_tokens": 106
},
"finish_reason": "stop",
"role": "assistant"
}
}
]
}
],
"$schema": "https://json.schemastore.org/aiconfig-1.0"
Expand Down

0 comments on commit a9784ad

Please sign in to comment.