mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-10-03 19:57:35 +00:00
1596 lines
43 KiB
YAML
1596 lines
43 KiB
YAML
components:
|
|
responses: {}
|
|
schemas:
|
|
AgenticSystemCreateRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
available_tools:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
input_shields:
|
|
items:
|
|
$ref: '#/components/schemas/ShieldConfig'
|
|
type: array
|
|
output_shields:
|
|
items:
|
|
$ref: '#/components/schemas/ShieldConfig'
|
|
type: array
|
|
parameters:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
oneOf:
|
|
- enum:
|
|
- web_search
|
|
- math
|
|
- image_gen
|
|
- code_interpreter
|
|
type: string
|
|
- type: string
|
|
required:
|
|
- tool_name
|
|
- input_shields
|
|
- output_shields
|
|
type: object
|
|
type: array
|
|
executable_tools:
|
|
items:
|
|
type: string
|
|
type: array
|
|
uniqueItems: true
|
|
input_shields:
|
|
items:
|
|
$ref: '#/components/schemas/ShieldConfig'
|
|
type: array
|
|
instructions:
|
|
type: string
|
|
model:
|
|
enum:
|
|
- llama3_8b_chat
|
|
- llama3_70b_chat
|
|
type: string
|
|
output_shields:
|
|
items:
|
|
$ref: '#/components/schemas/ShieldConfig'
|
|
type: array
|
|
required:
|
|
- instructions
|
|
- model
|
|
- available_tools
|
|
- executable_tools
|
|
- input_shields
|
|
- output_shields
|
|
type: object
|
|
AgenticSystemCreateResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
agent_id:
|
|
type: string
|
|
required:
|
|
- agent_id
|
|
type: object
|
|
AgenticSystemExecuteRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
agent_id:
|
|
type: string
|
|
messages:
|
|
items:
|
|
$ref: '#/components/schemas/Message'
|
|
type: array
|
|
stream:
|
|
default: false
|
|
type: boolean
|
|
turn_history:
|
|
items:
|
|
$ref: '#/components/schemas/AgenticSystemTurn'
|
|
type: array
|
|
required:
|
|
- agent_id
|
|
- messages
|
|
- turn_history
|
|
- stream
|
|
type: object
|
|
AgenticSystemExecuteResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
turn:
|
|
$ref: '#/components/schemas/AgenticSystemTurn'
|
|
required:
|
|
- turn
|
|
title: non-stream response from the agentic system.
|
|
type: object
|
|
AgenticSystemExecuteResponseStreamChunk:
|
|
additionalProperties: false
|
|
properties:
|
|
event_type:
|
|
enum:
|
|
- step_start
|
|
- step_end
|
|
- step_progress
|
|
title: The type of event.
|
|
type: string
|
|
response_text_delta:
|
|
type: string
|
|
retrieved_document:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
type: string
|
|
index_id:
|
|
type: string
|
|
required:
|
|
- index_id
|
|
- content
|
|
type: object
|
|
step_type:
|
|
enum:
|
|
- model_inference
|
|
- tool_execution
|
|
- safety_filtering
|
|
- memory_retrieval
|
|
title: The type of execution step.
|
|
type: string
|
|
step_uuid:
|
|
type: string
|
|
stop_reason:
|
|
enum:
|
|
- not_stopped
|
|
- finished_ok
|
|
- max_tokens
|
|
title: Stop reasons are used to indicate why the model stopped generating
|
|
text.
|
|
type: string
|
|
tool_call:
|
|
additionalProperties: false
|
|
properties:
|
|
arguments:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- arguments
|
|
title: A tool call is a request to a tool.
|
|
type: object
|
|
tool_response_delta:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- content
|
|
type: object
|
|
violation:
|
|
additionalProperties: false
|
|
properties:
|
|
details:
|
|
type: string
|
|
suggested_user_response:
|
|
type: string
|
|
violation_type:
|
|
type: string
|
|
required:
|
|
- violation_type
|
|
- details
|
|
type: object
|
|
required:
|
|
- event_type
|
|
- step_uuid
|
|
- step_type
|
|
title: Streamed agent execution response.
|
|
type: object
|
|
AgenticSystemTurn:
|
|
additionalProperties: false
|
|
properties:
|
|
response_message:
|
|
$ref: '#/components/schemas/Message'
|
|
steps:
|
|
items:
|
|
oneOf:
|
|
- additionalProperties: false
|
|
properties:
|
|
logprobs:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
step_type:
|
|
default: model_inference
|
|
enum:
|
|
- model_inference
|
|
- tool_execution
|
|
- safety_filtering
|
|
- memory_retrieval
|
|
title: The type of execution step.
|
|
type: string
|
|
text:
|
|
type: string
|
|
uuid:
|
|
type: string
|
|
required:
|
|
- step_type
|
|
- uuid
|
|
- text
|
|
type: object
|
|
- additionalProperties: false
|
|
properties:
|
|
step_type:
|
|
default: tool_execution
|
|
enum:
|
|
- model_inference
|
|
- tool_execution
|
|
- safety_filtering
|
|
- memory_retrieval
|
|
title: The type of execution step.
|
|
type: string
|
|
tool_calls:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
arguments:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- arguments
|
|
title: A tool call is a request to a tool.
|
|
type: object
|
|
type: array
|
|
tool_responses:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- content
|
|
type: object
|
|
type: array
|
|
uuid:
|
|
type: string
|
|
required:
|
|
- step_type
|
|
- uuid
|
|
- tool_calls
|
|
- tool_responses
|
|
type: object
|
|
- additionalProperties: false
|
|
properties:
|
|
step_type:
|
|
default: safety_filtering
|
|
enum:
|
|
- model_inference
|
|
- tool_execution
|
|
- safety_filtering
|
|
- memory_retrieval
|
|
title: The type of execution step.
|
|
type: string
|
|
uuid:
|
|
type: string
|
|
violation:
|
|
additionalProperties: false
|
|
properties:
|
|
details:
|
|
type: string
|
|
suggested_user_response:
|
|
type: string
|
|
violation_type:
|
|
type: string
|
|
required:
|
|
- violation_type
|
|
- details
|
|
type: object
|
|
required:
|
|
- step_type
|
|
- uuid
|
|
type: object
|
|
- additionalProperties: false
|
|
properties:
|
|
documents:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
type: string
|
|
index_id:
|
|
type: string
|
|
required:
|
|
- index_id
|
|
- content
|
|
type: object
|
|
type: array
|
|
scores:
|
|
items:
|
|
type: number
|
|
type: array
|
|
step_type:
|
|
default: memory_retrieval
|
|
enum:
|
|
- model_inference
|
|
- tool_execution
|
|
- safety_filtering
|
|
- memory_retrieval
|
|
title: The type of execution step.
|
|
type: string
|
|
uuid:
|
|
type: string
|
|
required:
|
|
- step_type
|
|
- uuid
|
|
- documents
|
|
- scores
|
|
type: object
|
|
type: array
|
|
user_messages:
|
|
items:
|
|
$ref: '#/components/schemas/Message'
|
|
type: array
|
|
required:
|
|
- user_messages
|
|
- steps
|
|
- response_message
|
|
title: A single turn in an interaction with an Agentic System.
|
|
type: object
|
|
Attachment:
|
|
additionalProperties: false
|
|
properties:
|
|
mime_type:
|
|
type: string
|
|
url:
|
|
$ref: '#/components/schemas/URL'
|
|
required:
|
|
- url
|
|
- mime_type
|
|
title: Attachments are used to refer to external resources, such as images,
|
|
videos, audio, etc.
|
|
type: object
|
|
ChatCompletionRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
available_tools:
|
|
items:
|
|
oneOf:
|
|
- enum:
|
|
- web_search
|
|
- math
|
|
- image_gen
|
|
- code_interpreter
|
|
type: string
|
|
- additionalProperties: false
|
|
properties:
|
|
input_shields:
|
|
items:
|
|
$ref: '#/components/schemas/ShieldConfig'
|
|
type: array
|
|
output_shields:
|
|
items:
|
|
$ref: '#/components/schemas/ShieldConfig'
|
|
type: array
|
|
parameters:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
oneOf:
|
|
- enum:
|
|
- web_search
|
|
- math
|
|
- image_gen
|
|
- code_interpreter
|
|
type: string
|
|
- type: string
|
|
required:
|
|
- tool_name
|
|
- input_shields
|
|
- output_shields
|
|
type: object
|
|
type: array
|
|
logprobs:
|
|
default: false
|
|
type: boolean
|
|
max_tokens:
|
|
default: 0
|
|
type: integer
|
|
message:
|
|
$ref: '#/components/schemas/Message'
|
|
message_history:
|
|
items:
|
|
$ref: '#/components/schemas/Message'
|
|
type: array
|
|
model:
|
|
enum:
|
|
- llama3_8b_chat
|
|
- llama3_70b_chat
|
|
type: string
|
|
sampling_params:
|
|
additionalProperties: false
|
|
properties:
|
|
strategy:
|
|
default: greedy
|
|
type: string
|
|
temperature:
|
|
default: 0.0
|
|
type: number
|
|
top_k:
|
|
default: 0
|
|
type: integer
|
|
top_p:
|
|
default: 0.95
|
|
type: number
|
|
required:
|
|
- temperature
|
|
- strategy
|
|
- top_p
|
|
- top_k
|
|
type: object
|
|
stream:
|
|
default: false
|
|
type: boolean
|
|
required:
|
|
- message
|
|
- model
|
|
- message_history
|
|
- sampling_params
|
|
- available_tools
|
|
- max_tokens
|
|
- stream
|
|
- logprobs
|
|
type: object
|
|
ChatCompletionResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
logprobs:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
stop_reason:
|
|
enum:
|
|
- not_stopped
|
|
- finished_ok
|
|
- max_tokens
|
|
title: Stop reasons are used to indicate why the model stopped generating
|
|
text.
|
|
type: string
|
|
tool_calls:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
arguments:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- arguments
|
|
title: A tool call is a request to a tool.
|
|
type: object
|
|
type: array
|
|
required:
|
|
- content
|
|
- tool_calls
|
|
title: Normal chat completion response.
|
|
type: object
|
|
ChatCompletionResponseStreamChunk:
|
|
additionalProperties: false
|
|
properties:
|
|
stop_reason:
|
|
enum:
|
|
- not_stopped
|
|
- finished_ok
|
|
- max_tokens
|
|
title: Stop reasons are used to indicate why the model stopped generating
|
|
text.
|
|
type: string
|
|
text_delta:
|
|
type: string
|
|
tool_call:
|
|
additionalProperties: false
|
|
properties:
|
|
arguments:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- arguments
|
|
title: A tool call is a request to a tool.
|
|
type: object
|
|
required:
|
|
- text_delta
|
|
title: Streamed chat completion response. The actual response is a series of
|
|
such objects.
|
|
type: object
|
|
CompletionRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
logprobs:
|
|
default: false
|
|
type: boolean
|
|
max_tokens:
|
|
default: 0
|
|
type: integer
|
|
model:
|
|
enum:
|
|
- llama3_8b
|
|
- llama3_70b
|
|
type: string
|
|
sampling_params:
|
|
additionalProperties: false
|
|
properties:
|
|
strategy:
|
|
default: greedy
|
|
type: string
|
|
temperature:
|
|
default: 0.0
|
|
type: number
|
|
top_k:
|
|
default: 0
|
|
type: integer
|
|
top_p:
|
|
default: 0.95
|
|
type: number
|
|
required:
|
|
- temperature
|
|
- strategy
|
|
- top_p
|
|
- top_k
|
|
type: object
|
|
stream:
|
|
default: false
|
|
type: boolean
|
|
required:
|
|
- content
|
|
- model
|
|
- sampling_params
|
|
- max_tokens
|
|
- stream
|
|
- logprobs
|
|
type: object
|
|
CompletionResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
logprobs:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
stop_reason:
|
|
enum:
|
|
- not_stopped
|
|
- finished_ok
|
|
- max_tokens
|
|
title: Stop reasons are used to indicate why the model stopped generating
|
|
text.
|
|
type: string
|
|
required:
|
|
- content
|
|
title: Normal completion response.
|
|
type: object
|
|
CompletionResponseStreamChunk:
|
|
additionalProperties: false
|
|
properties:
|
|
logprobs:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
stop_reason:
|
|
enum:
|
|
- not_stopped
|
|
- finished_ok
|
|
- max_tokens
|
|
title: Stop reasons are used to indicate why the model stopped generating
|
|
text.
|
|
type: string
|
|
text_delta:
|
|
type: string
|
|
required:
|
|
- text_delta
|
|
title: streamed completion response.
|
|
type: object
|
|
CreateDatasetRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
dataset:
|
|
$ref: '#/components/schemas/Dataset'
|
|
uuid:
|
|
type: string
|
|
required:
|
|
- uuid
|
|
- dataset
|
|
title: Request to create a dataset.
|
|
type: object
|
|
Dataset:
|
|
additionalProperties: false
|
|
properties:
|
|
columns:
|
|
additionalProperties:
|
|
enum:
|
|
- dialog
|
|
- text
|
|
- media
|
|
- number
|
|
- json
|
|
type: string
|
|
type: object
|
|
content_url:
|
|
$ref: '#/components/schemas/URL'
|
|
metadata:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
required:
|
|
- columns
|
|
- content_url
|
|
- metadata
|
|
title: Dataset to be used for training or evaluating language models.
|
|
type: object
|
|
FinetuningJobLogStream:
|
|
additionalProperties: false
|
|
properties:
|
|
job_uuid:
|
|
type: string
|
|
log_lines:
|
|
items:
|
|
type: string
|
|
type: array
|
|
required:
|
|
- job_uuid
|
|
- log_lines
|
|
title: Stream of logs from a finetuning job.
|
|
type: object
|
|
FinetuningJobStatusResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
completed_at:
|
|
format: date-time
|
|
type: string
|
|
job_uuid:
|
|
type: string
|
|
resources_allocated:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
scheduled_at:
|
|
format: date-time
|
|
type: string
|
|
started_at:
|
|
format: date-time
|
|
type: string
|
|
status:
|
|
enum:
|
|
- running
|
|
- completed
|
|
- failed
|
|
- scheduled
|
|
type: string
|
|
required:
|
|
- job_uuid
|
|
- status
|
|
title: Status of a finetuning job.
|
|
type: object
|
|
FinetuningTrainRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
algorithm:
|
|
enum:
|
|
- full
|
|
- lora
|
|
- qlora
|
|
- dora
|
|
type: string
|
|
algorithm_config:
|
|
oneOf:
|
|
- $ref: '#/components/schemas/LoraFinetuningConfig'
|
|
- additionalProperties: false
|
|
properties:
|
|
alpha:
|
|
type: integer
|
|
apply_lora_to_mlp:
|
|
type: boolean
|
|
apply_lora_to_output:
|
|
type: boolean
|
|
lora_attn_modules:
|
|
items:
|
|
type: string
|
|
type: array
|
|
rank:
|
|
type: integer
|
|
required:
|
|
- lora_attn_modules
|
|
- apply_lora_to_mlp
|
|
- apply_lora_to_output
|
|
- rank
|
|
- alpha
|
|
type: object
|
|
- additionalProperties: false
|
|
properties:
|
|
alpha:
|
|
type: integer
|
|
apply_lora_to_mlp:
|
|
type: boolean
|
|
apply_lora_to_output:
|
|
type: boolean
|
|
lora_attn_modules:
|
|
items:
|
|
type: string
|
|
type: array
|
|
rank:
|
|
type: integer
|
|
required:
|
|
- lora_attn_modules
|
|
- apply_lora_to_mlp
|
|
- apply_lora_to_output
|
|
- rank
|
|
- alpha
|
|
type: object
|
|
dataset:
|
|
$ref: '#/components/schemas/Dataset'
|
|
hyperparam_search_config:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
job_uuid:
|
|
type: string
|
|
logger_config:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
model:
|
|
enum:
|
|
- llama3_8b
|
|
- llama3_70b
|
|
type: string
|
|
optimizer_config:
|
|
$ref: '#/components/schemas/OptimizerConfig'
|
|
training_config:
|
|
$ref: '#/components/schemas/TrainingConfig'
|
|
validation_dataset:
|
|
$ref: '#/components/schemas/Dataset'
|
|
required:
|
|
- job_uuid
|
|
- model
|
|
- dataset
|
|
- validation_dataset
|
|
- algorithm
|
|
- algorithm_config
|
|
- optimizer_config
|
|
- training_config
|
|
- hyperparam_search_config
|
|
- logger_config
|
|
title: Request to finetune a model.
|
|
type: object
|
|
KScoredPromptGenerations:
|
|
additionalProperties: false
|
|
properties:
|
|
k_scored_generations:
|
|
items:
|
|
$ref: '#/components/schemas/MessageScore'
|
|
type: array
|
|
prompt:
|
|
$ref: '#/components/schemas/Message'
|
|
required:
|
|
- prompt
|
|
- k_scored_generations
|
|
type: object
|
|
LoraFinetuningConfig:
|
|
additionalProperties: false
|
|
properties:
|
|
alpha:
|
|
type: integer
|
|
apply_lora_to_mlp:
|
|
type: boolean
|
|
apply_lora_to_output:
|
|
type: boolean
|
|
lora_attn_modules:
|
|
items:
|
|
type: string
|
|
type: array
|
|
rank:
|
|
type: integer
|
|
required:
|
|
- lora_attn_modules
|
|
- apply_lora_to_mlp
|
|
- apply_lora_to_output
|
|
- rank
|
|
- alpha
|
|
type: object
|
|
Message:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
role:
|
|
enum:
|
|
- system
|
|
- user
|
|
- assistant
|
|
- tool
|
|
type: string
|
|
tool_calls:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
arguments:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- arguments
|
|
title: A tool call is a request to a tool.
|
|
type: object
|
|
type: array
|
|
tool_responses:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
content:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
- items:
|
|
oneOf:
|
|
- type: string
|
|
- $ref: '#/components/schemas/Attachment'
|
|
type: array
|
|
tool_name:
|
|
type: string
|
|
required:
|
|
- tool_name
|
|
- content
|
|
type: object
|
|
type: array
|
|
required:
|
|
- role
|
|
- content
|
|
- tool_calls
|
|
- tool_responses
|
|
type: object
|
|
MessageScore:
|
|
additionalProperties: false
|
|
properties:
|
|
message:
|
|
$ref: '#/components/schemas/Message'
|
|
score:
|
|
type: number
|
|
required:
|
|
- message
|
|
- score
|
|
title: A single message and its score.
|
|
type: object
|
|
OptimizerConfig:
|
|
additionalProperties: false
|
|
properties:
|
|
lr:
|
|
type: number
|
|
lr_min:
|
|
type: number
|
|
optimizer_type:
|
|
enum:
|
|
- adam
|
|
- adamw
|
|
- sgd
|
|
type: string
|
|
weight_decay:
|
|
type: number
|
|
required:
|
|
- optimizer_type
|
|
- lr
|
|
- lr_min
|
|
- weight_decay
|
|
type: object
|
|
RewardScoringRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
model:
|
|
type: string
|
|
prompt_generations:
|
|
items:
|
|
additionalProperties: false
|
|
properties:
|
|
k_generations:
|
|
items:
|
|
$ref: '#/components/schemas/Message'
|
|
type: array
|
|
message_history:
|
|
items:
|
|
$ref: '#/components/schemas/Message'
|
|
type: array
|
|
prompt:
|
|
$ref: '#/components/schemas/Message'
|
|
required:
|
|
- prompt
|
|
- message_history
|
|
- k_generations
|
|
type: object
|
|
type: array
|
|
required:
|
|
- prompt_generations
|
|
- model
|
|
title: Request to score a reward function. A list of prompts and a list of responses
|
|
per prompt.
|
|
type: object
|
|
RewardScoringResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
scored_generations:
|
|
items:
|
|
$ref: '#/components/schemas/KScoredPromptGenerations'
|
|
type: array
|
|
required:
|
|
- scored_generations
|
|
title: Response from the reward scoring. Batch of (prompt, response, score)
|
|
tuples that pass the threshold.
|
|
type: object
|
|
ShieldConfig:
|
|
additionalProperties: false
|
|
properties:
|
|
params:
|
|
additionalProperties:
|
|
oneOf:
|
|
- type: 'null'
|
|
- type: boolean
|
|
- type: number
|
|
- type: string
|
|
- type: array
|
|
- type: object
|
|
type: object
|
|
shield_type:
|
|
enum:
|
|
- llama_guard
|
|
- prompt_guard
|
|
- code_guard
|
|
title: The type of safety shield.
|
|
type: string
|
|
required:
|
|
- shield_type
|
|
- params
|
|
type: object
|
|
SyntheticDataGenerationRequest:
|
|
additionalProperties: false
|
|
properties:
|
|
filtering_function:
|
|
default: none
|
|
enum:
|
|
- none
|
|
- random
|
|
- top_k
|
|
- top_p
|
|
- top_k_top_p
|
|
- sigmoid
|
|
title: The type of filtering function.
|
|
type: string
|
|
prompts:
|
|
items:
|
|
type: string
|
|
type: array
|
|
required:
|
|
- prompts
|
|
- filtering_function
|
|
title: Request to generate synthetic data. A small batch of prompts and a filtering
|
|
function
|
|
type: object
|
|
SyntheticDataGenerationResponse:
|
|
additionalProperties: false
|
|
properties:
|
|
statistics:
|
|
additionalProperties:
|
|
type: number
|
|
type: object
|
|
synthetic_data:
|
|
items:
|
|
maxItems: 3
|
|
minItems: 3
|
|
prefixItems:
|
|
- type: string
|
|
- type: string
|
|
- type: number
|
|
type: array
|
|
type: array
|
|
required:
|
|
- synthetic_data
|
|
- statistics
|
|
title: Response from the synthetic data generation. Batch of (prompt, response,
|
|
score) tuples that pass the threshold.
|
|
type: object
|
|
TrainingConfig:
|
|
additionalProperties: false
|
|
properties:
|
|
batch_size:
|
|
type: integer
|
|
enable_activation_checkpointing:
|
|
type: boolean
|
|
fsdp_cpu_offload:
|
|
type: boolean
|
|
memory_efficient_fsdp_wrap:
|
|
type: boolean
|
|
n_epochs:
|
|
type: integer
|
|
n_iters:
|
|
type: integer
|
|
shuffle:
|
|
type: boolean
|
|
required:
|
|
- n_epochs
|
|
- batch_size
|
|
- shuffle
|
|
- n_iters
|
|
- enable_activation_checkpointing
|
|
- memory_efficient_fsdp_wrap
|
|
- fsdp_cpu_offload
|
|
type: object
|
|
URL:
|
|
format: uri
|
|
pattern: ^(https?://|file://|data:)
|
|
type: string
|
|
info:
|
|
description: This is the llama stack
|
|
title: Llama Stack specification
|
|
version: '0.1'
|
|
jsonSchemaDialect: https://json-schema.org/draft/2020-12/schema
|
|
openapi: 3.1.0
|
|
paths:
|
|
/agentic_system/create:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/AgenticSystemCreateRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/AgenticSystemCreateResponse'
|
|
description: OK
|
|
tags:
|
|
- AgenticSystem
|
|
/agentic_system/delete:
|
|
delete:
|
|
parameters:
|
|
- in: query
|
|
name: agent_id
|
|
required: true
|
|
schema:
|
|
type: string
|
|
responses:
|
|
'200':
|
|
description: OK
|
|
tags:
|
|
- AgenticSystem
|
|
/agentic_system/execute:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/AgenticSystemExecuteRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
oneOf:
|
|
- $ref: '#/components/schemas/AgenticSystemExecuteResponse'
|
|
- $ref: '#/components/schemas/AgenticSystemExecuteResponseStreamChunk'
|
|
description: non-stream response from the agentic system. **OR** Streamed
|
|
agent execution response.
|
|
tags:
|
|
- AgenticSystem
|
|
/chat_completion:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/ChatCompletionRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
oneOf:
|
|
- $ref: '#/components/schemas/ChatCompletionResponse'
|
|
- $ref: '#/components/schemas/ChatCompletionResponseStreamChunk'
|
|
description: Normal chat completion response. **OR** Streamed chat completion
|
|
response. The actual response is a series of such objects.
|
|
tags:
|
|
- Inference
|
|
/completion:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/CompletionRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
oneOf:
|
|
- $ref: '#/components/schemas/CompletionResponse'
|
|
- $ref: '#/components/schemas/CompletionResponseStreamChunk'
|
|
description: Normal completion response. **OR** streamed completion response.
|
|
tags:
|
|
- Inference
|
|
/datasets/create:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/CreateDatasetRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
description: OK
|
|
tags:
|
|
- Datasets
|
|
/datasets/delete:
|
|
delete:
|
|
parameters:
|
|
- in: query
|
|
name: dataset_id
|
|
required: true
|
|
schema:
|
|
type: string
|
|
responses:
|
|
'200':
|
|
description: OK
|
|
tags:
|
|
- Datasets
|
|
/datasets/get:
|
|
get:
|
|
parameters:
|
|
- in: query
|
|
name: dataset_id
|
|
required: true
|
|
schema:
|
|
type: string
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/Dataset'
|
|
description: OK
|
|
tags:
|
|
- Datasets
|
|
/finetuning/job/logs:
|
|
get:
|
|
parameters:
|
|
- in: query
|
|
name: job_uuid
|
|
required: true
|
|
schema:
|
|
type: string
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/FinetuningJobLogStream'
|
|
description: OK
|
|
tags:
|
|
- Finetuning
|
|
/finetuning/job/status:
|
|
get:
|
|
parameters:
|
|
- in: query
|
|
name: job_uuid
|
|
required: true
|
|
schema:
|
|
type: string
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/FinetuningJobStatusResponse'
|
|
description: OK
|
|
tags:
|
|
- Finetuning
|
|
/finetuning/text_generation/train:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/FinetuningTrainRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
description: OK
|
|
tags:
|
|
- Finetuning
|
|
/reward_scoring/score:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/RewardScoringRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/RewardScoringResponse'
|
|
description: OK
|
|
tags:
|
|
- RewardScoring
|
|
/synthetic_data_generation/generate:
|
|
post:
|
|
parameters: []
|
|
requestBody:
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/SyntheticDataGenerationRequest'
|
|
required: true
|
|
responses:
|
|
'200':
|
|
content:
|
|
application/json:
|
|
schema:
|
|
$ref: '#/components/schemas/SyntheticDataGenerationResponse'
|
|
description: OK
|
|
tags:
|
|
- SyntheticDataGeneration
|
|
security:
|
|
- Default: []
|
|
servers:
|
|
- url: http://llama.meta.com
|
|
tags:
|
|
- name: Inference
|
|
- name: SyntheticDataGeneration
|
|
- name: Datasets
|
|
- name: AgenticSystem
|
|
- name: Finetuning
|
|
- name: RewardScoring
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/ShieldConfig" />
|
|
name: ShieldConfig
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/AgenticSystemCreateRequest"
|
|
/>
|
|
name: AgenticSystemCreateRequest
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/AgenticSystemCreateResponse"
|
|
/>
|
|
name: AgenticSystemCreateResponse
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/AgenticSystemExecuteRequest"
|
|
/>
|
|
name: AgenticSystemExecuteRequest
|
|
- description: 'A single turn in an interaction with an Agentic System.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/AgenticSystemTurn" />'
|
|
name: AgenticSystemTurn
|
|
- description: 'Attachments are used to refer to external resources, such as images,
|
|
videos, audio, etc.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/Attachment" />'
|
|
name: Attachment
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/Message" />
|
|
name: Message
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/URL" />
|
|
name: URL
|
|
- description: 'non-stream response from the agentic system.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/AgenticSystemExecuteResponse"
|
|
/>'
|
|
name: AgenticSystemExecuteResponse
|
|
- description: 'Streamed agent execution response.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/AgenticSystemExecuteResponseStreamChunk"
|
|
/>'
|
|
name: AgenticSystemExecuteResponseStreamChunk
|
|
- description: 'Request to create a dataset.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/CreateDatasetRequest" />'
|
|
name: CreateDatasetRequest
|
|
- description: 'Dataset to be used for training or evaluating language models.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/Dataset" />'
|
|
name: Dataset
|
|
- description: 'Status of a finetuning job.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/FinetuningJobStatusResponse"
|
|
/>'
|
|
name: FinetuningJobStatusResponse
|
|
- description: 'Stream of logs from a finetuning job.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/FinetuningJobLogStream" />'
|
|
name: FinetuningJobLogStream
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/ChatCompletionRequest"
|
|
/>
|
|
name: ChatCompletionRequest
|
|
- description: 'Normal chat completion response.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/ChatCompletionResponse" />'
|
|
name: ChatCompletionResponse
|
|
- description: 'Streamed chat completion response. The actual response is a series
|
|
of such objects.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/ChatCompletionResponseStreamChunk"
|
|
/>'
|
|
name: ChatCompletionResponseStreamChunk
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/CompletionRequest"
|
|
/>
|
|
name: CompletionRequest
|
|
- description: 'Normal completion response.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/CompletionResponse" />'
|
|
name: CompletionResponse
|
|
- description: 'streamed completion response.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/CompletionResponseStreamChunk"
|
|
/>'
|
|
name: CompletionResponseStreamChunk
|
|
- description: 'Request to generate synthetic data. A small batch of prompts and a
|
|
filtering function
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/SyntheticDataGenerationRequest"
|
|
/>'
|
|
name: SyntheticDataGenerationRequest
|
|
- description: 'Response from the synthetic data generation. Batch of (prompt, response,
|
|
score) tuples that pass the threshold.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/SyntheticDataGenerationResponse"
|
|
/>'
|
|
name: SyntheticDataGenerationResponse
|
|
- description: 'Request to score a reward function. A list of prompts and a list of
|
|
responses per prompt.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/RewardScoringRequest" />'
|
|
name: RewardScoringRequest
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/KScoredPromptGenerations"
|
|
/>
|
|
name: KScoredPromptGenerations
|
|
- description: 'A single message and its score.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/MessageScore" />'
|
|
name: MessageScore
|
|
- description: 'Response from the reward scoring. Batch of (prompt, response, score)
|
|
tuples that pass the threshold.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/RewardScoringResponse" />'
|
|
name: RewardScoringResponse
|
|
- description: 'Request to finetune a model.
|
|
|
|
|
|
<SchemaDefinition schemaRef="#/components/schemas/FinetuningTrainRequest" />'
|
|
name: FinetuningTrainRequest
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/LoraFinetuningConfig"
|
|
/>
|
|
name: LoraFinetuningConfig
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/OptimizerConfig"
|
|
/>
|
|
name: OptimizerConfig
|
|
- description: <SchemaDefinition schemaRef="#/components/schemas/TrainingConfig" />
|
|
name: TrainingConfig
|
|
x-tagGroups:
|
|
- name: Operations
|
|
tags:
|
|
- AgenticSystem
|
|
- Datasets
|
|
- Finetuning
|
|
- Inference
|
|
- RewardScoring
|
|
- SyntheticDataGeneration
|
|
- name: Types
|
|
tags:
|
|
- AgenticSystemCreateRequest
|
|
- AgenticSystemCreateResponse
|
|
- AgenticSystemExecuteRequest
|
|
- AgenticSystemExecuteResponse
|
|
- AgenticSystemExecuteResponseStreamChunk
|
|
- AgenticSystemTurn
|
|
- Attachment
|
|
- ChatCompletionRequest
|
|
- ChatCompletionResponse
|
|
- ChatCompletionResponseStreamChunk
|
|
- CompletionRequest
|
|
- CompletionResponse
|
|
- CompletionResponseStreamChunk
|
|
- CreateDatasetRequest
|
|
- Dataset
|
|
- FinetuningJobLogStream
|
|
- FinetuningJobStatusResponse
|
|
- FinetuningTrainRequest
|
|
- KScoredPromptGenerations
|
|
- LoraFinetuningConfig
|
|
- Message
|
|
- MessageScore
|
|
- OptimizerConfig
|
|
- RewardScoringRequest
|
|
- RewardScoringResponse
|
|
- ShieldConfig
|
|
- SyntheticDataGenerationRequest
|
|
- SyntheticDataGenerationResponse
|
|
- TrainingConfig
|
|
- URL
|