phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-04 04:04:14 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fixes to reward stuff

Browse source
This commit is contained in:
Ashwin Bharambe 2024-07-10 19:22:33 -07:00
parent eb12bfbef0
commit 956f07b04c
3 changed files with 461 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

247
source/openapi.html
View file

@ -191,6 +191,66 @@
"required": true
}
}
},
"/synthetic_data_generation/generate": {
"post": {
"responses": {
"200": {
"description": "OK",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/SyntheticDataGenerationResponse"
}
}
}
}
},
"tags": [
"SyntheticDataGeneration"
],
"parameters": [],
"requestBody": {
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/SyntheticDataGenerationRequest"
}
}
},
"required": true
}
}
},
"/reward_scoring/score": {
"post": {
"responses": {
"200": {
"description": "OK",
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/RewardScoringResponse"
}
}
}
}
},
"tags": [
"RewardScoring"
],
"parameters": [],
"requestBody": {
"content": {
"application/json": {
"schema": {
"$ref": "#/components/schemas/RewardScoringRequest"
}
}
},
"required": true
}
}
}
},
"jsonSchemaDialect": "https://json-schema.org/draft/2020-12/schema",
@ -1451,6 +1511,161 @@
"text_delta"
],
"title": "streamed completion response."
},
"SyntheticDataGenerationRequest": {
"type": "object",
"properties": {
"prompts": {
"type": "array",
"items": {
"type": "string"
}
},
"filtering_function": {
"type": "string",
"enum": [
"none",
"random",
"top_k",
"top_p",
"top_k_top_p",
"sigmoid"
],
"title": "The type of filtering function.",
"default": "none"
}
},
"additionalProperties": false,
"required": [
"prompts",
"filtering_function"
],
"title": "Request to generate synthetic data. A small batch of prompts and a filtering function"
},
"SyntheticDataGenerationResponse": {
"type": "object",
"properties": {
"synthetic_data": {
"type": "array",
"items": {
"type": "array",
"minItems": 3,
"maxItems": 3,
"prefixItems": [
{
"type": "string"
},
{
"type": "string"
},
{
"type": "number"
}
]
}
},
"statistics": {
"type": "object",
"additionalProperties": {
"type": "number"
}
}
},
"additionalProperties": false,
"required": [
"synthetic_data",
"statistics"
],
"title": "Response from the synthetic data generation. Batch of (prompt, response, score) tuples that pass the threshold."
},
"RewardScoringRequest": {
"type": "object",
"properties": {
"prompt_generations": {
"type": "array",
"items": {
"type": "object",
"properties": {
"prompt": {
"$ref": "#/components/schemas/Message"
},
"message_history": {
"type": "array",
"items": {
"$ref": "#/components/schemas/Message"
}
},
"generation": {
"$ref": "#/components/schemas/Message"
}
},
"additionalProperties": false,
"required": [
"prompt",
"message_history",
"generation"
]
}
},
"model": {
"type": "string"
}
},
"additionalProperties": false,
"required": [
"prompt_generations",
"model"
],
"title": "Request to score a reward function. A list of prompts and a list of responses per prompt."
},
"RewardScoringResponse": {
"type": "object",
"properties": {
"scored_generations": {
"type": "array",
"items": {
"type": "object",
"properties": {
"prompt_generation": {
"type": "object",
"properties": {
"prompt": {
"$ref": "#/components/schemas/Message"
},
"message_history": {
"type": "array",
"items": {
"$ref": "#/components/schemas/Message"
}
},
"generation": {
"$ref": "#/components/schemas/Message"
}
},
"additionalProperties": false,
"required": [
"prompt",
"message_history",
"generation"
]
},
"score": {
"type": "number"
}
},
"additionalProperties": false,
"required": [
"prompt_generation",
"score"
]
}
}
},
"additionalProperties": false,
"required": [
"scored_generations"
],
"title": "Response from the reward scoring. Batch of (prompt, response, score) tuples that pass the threshold."
}
},
"responses": {}
@ -1462,11 +1677,17 @@
],
"tags": [
{
"name": "AgenticSystem"
"name": "RewardScoring"
},
{
"name": "Inference"
},
{
"name": "SyntheticDataGeneration"
},
{
"name": "AgenticSystem"
},
{
"name": "ShieldConfig",
"description": "<SchemaDefinition schemaRef=\"#/components/schemas/ShieldConfig\" />"
@ -1530,6 +1751,22 @@
{
"name": "CompletionResponseStreamChunk",
"description": "streamed completion response.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/CompletionResponseStreamChunk\" />"
},
{
"name": "SyntheticDataGenerationRequest",
"description": "Request to generate synthetic data. A small batch of prompts and a filtering function\n\n<SchemaDefinition schemaRef=\"#/components/schemas/SyntheticDataGenerationRequest\" />"
},
{
"name": "SyntheticDataGenerationResponse",
"description": "Response from the synthetic data generation. Batch of (prompt, response, score) tuples that pass the threshold.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/SyntheticDataGenerationResponse\" />"
},
{
"name": "RewardScoringRequest",
"description": "Request to score a reward function. A list of prompts and a list of responses per prompt.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/RewardScoringRequest\" />"
},
{
"name": "RewardScoringResponse",
"description": "Response from the reward scoring. Batch of (prompt, response, score) tuples that pass the threshold.\n\n<SchemaDefinition schemaRef=\"#/components/schemas/RewardScoringResponse\" />"
}
],
"x-tagGroups": [
@ -1537,7 +1774,9 @@
"name": "Operations",
"tags": [
"AgenticSystem",
"Inference"
"Inference",
"RewardScoring",
"SyntheticDataGeneration"
]
},
{
@ -1557,7 +1796,11 @@
"CompletionResponse",
"CompletionResponseStreamChunk",
"Message",
"RewardScoringRequest",
"RewardScoringResponse",
"ShieldConfig",
"SyntheticDataGenerationRequest",
"SyntheticDataGenerationResponse",
"URL"
]
}