phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-07-29 15:23:51 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

added more docs

Browse source
This commit is contained in:
Raghotham Murthy 2024-07-11 03:10:30 -07:00
parent ab44e9c862
commit e657e71446
3 changed files with 51 additions and 54 deletions
Download patch file Download diff file Expand all files Collapse all files

4
source/api_definitions.py
View file

@ -550,9 +550,7 @@ if __name__ == "__main__":
info=Info( info=Info(
title="[DRAFT] Llama Stack Specification", title="[DRAFT] Llama Stack Specification",
version="0.0.1", version="0.0.1",
description=""" description="""Meta has built out a fairly sophisticated platform internally to post train, evaluate, and
Meta has built out a fairly sophisticated platform internally to post train, evaluate, and
serve Llama models to support Meta’s products. Given the newer capabilities of the llama models, serve Llama models to support Meta’s products. Given the newer capabilities of the llama models,
the model development and model serving capabilities of the platform need to be enhanced in the model development and model serving capabilities of the platform need to be enhanced in
specific ways in order to best leverage the models. For example, the inference platform needs specific ways in order to best leverage the models. For example, the inference platform needs

24
source/openapi.html
View file

@ -21,7 +21,7 @@
"info": { "info": {
"title": "[DRAFT] Llama Stack Specification", "title": "[DRAFT] Llama Stack Specification",
"version": "0.0.1", "version": "0.0.1",
"description": "\n \n Meta has built out a fairly sophisticated platform internally to post train, evaluate, and \n serve Llama models to support Meta’s products. Given the newer capabilities of the llama models, \n the model development and model serving capabilities of the platform need to be enhanced in \n specific ways in order to best leverage the models. For example, the inference platform needs \n to support code execution to take advantage of the built-in knowledge of tools of the model. \n The largest models are of high enough quality to be used to generate synthetic data or be used \n as reward models. There are specific fine tuning and quantization techniques that we have found \n result in the best performing Llama models. We would like to share ways in which an LLM Ops \n toolchain can be designed by leveraging our learnings in getting Llama models to power Meta’s products.\n\n In addition, the Llama 3 models Meta will release in July should not just be seen as a model, but \n really as a system starting the transition towards an entity capable of performing \"agentic\" tasks \n which require the ability to act as the central planner and break a task down and perform multi-step \n reasoning and call tools for specific operations. In addition, there needs to be general model-level \n safety checks as well as task-specific safety checks that are performed at a system level. \n\n We are defining the Llama Stack as a set of APIs and standards by synthesizing our learnings while \n working with Llama models. The APIs are divided into the llama-toolchain-api and the llama-agentic-system-api. \n These APIs provide a coherent way for model developers to fine tune and serve Llama models, and agentic app \n developers to leverage all the capabilities of the Llama models seamlessly. We would like to work with the \n ecosystem to enhance and simplify the API. In addition, we will be releasing a plug-in architecture to allow \n creating distributions of the llama stack with different implementations.\n\n\n This is the specification of the llama stack that provides \n a set of endpoints and their corresponding interfaces that are tailored to \n best leverage Llama Models. The specification is still in draft and subject to change." "description": "Meta has built out a fairly sophisticated platform internally to post train, evaluate, and \n serve Llama models to support Meta’s products. Given the newer capabilities of the llama models, \n the model development and model serving capabilities of the platform need to be enhanced in \n specific ways in order to best leverage the models. For example, the inference platform needs \n to support code execution to take advantage of the built-in knowledge of tools of the model. \n The largest models are of high enough quality to be used to generate synthetic data or be used \n as reward models. There are specific fine tuning and quantization techniques that we have found \n result in the best performing Llama models. We would like to share ways in which an LLM Ops \n toolchain can be designed by leveraging our learnings in getting Llama models to power Meta’s products.\n\n In addition, the Llama 3 models Meta will release in July should not just be seen as a model, but \n really as a system starting the transition towards an entity capable of performing \"agentic\" tasks \n which require the ability to act as the central planner and break a task down and perform multi-step \n reasoning and call tools for specific operations. In addition, there needs to be general model-level \n safety checks as well as task-specific safety checks that are performed at a system level. \n\n We are defining the Llama Stack as a set of APIs and standards by synthesizing our learnings while \n working with Llama models. The APIs are divided into the llama-toolchain-api and the llama-agentic-system-api. \n These APIs provide a coherent way for model developers to fine tune and serve Llama models, and agentic app \n developers to leverage all the capabilities of the Llama models seamlessly. We would like to work with the \n ecosystem to enhance and simplify the API. In addition, we will be releasing a plug-in architecture to allow \n creating distributions of the llama stack with different implementations.\n\n\n This is the specification of the llama stack that provides \n a set of endpoints and their corresponding interfaces that are tailored to \n best leverage Llama Models. The specification is still in draft and subject to change."
}, },
"servers": [ "servers": [
{ {
@ -3331,26 +3331,26 @@
} }
], ],
"tags": [ "tags": [
{
"name": "PostTraining"
},
{
"name": "Inference"
},
{ {
"name": "MemoryBanks" "name": "MemoryBanks"
}, },
{ {
"name": "SyntheticDataGeneration" "name": "Datasets"
},
{
"name": "RewardScoring"
}, },
{ {
"name": "AgenticSystem" "name": "AgenticSystem"
}, },
{ {
"name": "Datasets" "name": "SyntheticDataGeneration"
},
{
"name": "PostTraining"
},
{
"name": "RewardScoring"
},
{
"name": "Inference"
}, },
{ {
"name": "ShieldConfig", "name": "ShieldConfig",

77
source/openapi.yaml
View file

@ -1502,40 +1502,39 @@ components:
pattern: ^(https?://|file://|data:) pattern: ^(https?://|file://|data:)
type: string type: string
info: info:
description: "\n \n Meta has built out a fairly sophisticated\ description: "Meta has built out a fairly sophisticated platform internally to post\
\ platform internally to post train, evaluate, and \n serve Llama\ \ train, evaluate, and \n serve Llama models to support Meta’s\
\ models to support Meta’s products. Given the newer capabilities of the llama\ \ products. Given the newer capabilities of the llama models, \n \
\ models, \n the model development and model serving capabilities\ \ the model development and model serving capabilities of the platform need\
\ of the platform need to be enhanced in \n specific ways in order\ \ to be enhanced in \n specific ways in order to best leverage\
\ to best leverage the models. For example, the inference platform needs \n \ \ the models. For example, the inference platform needs \n to support\
\ to support code execution to take advantage of the built-in knowledge\ \ code execution to take advantage of the built-in knowledge of tools of the model.\
\ of tools of the model. \n The largest models are of high enough\ \ \n The largest models are of high enough quality to be used to\
\ quality to be used to generate synthetic data or be used \n as\ \ generate synthetic data or be used \n as reward models. There\
\ reward models. There are specific fine tuning and quantization techniques that\ \ are specific fine tuning and quantization techniques that we have found \n \
\ we have found \n result in the best performing Llama models.\ \ result in the best performing Llama models. We would like to share\
\ We would like to share ways in which an LLM Ops \n toolchain\ \ ways in which an LLM Ops \n toolchain can be designed by leveraging\
\ can be designed by leveraging our learnings in getting Llama models to power\ \ our learnings in getting Llama models to power Meta’s products.\n\n \
\ Meta’s products.\n\n In addition, the Llama 3 models Meta will\ \ In addition, the Llama 3 models Meta will release in July should not\
\ release in July should not just be seen as a model, but \n really\ \ just be seen as a model, but \n really as a system starting the\
\ as a system starting the transition towards an entity capable of performing\ \ transition towards an entity capable of performing \"agentic\" tasks \n \
\ \"agentic\" tasks \n which require the ability to act as the\ \ which require the ability to act as the central planner and break\
\ central planner and break a task down and perform multi-step \n \ \ a task down and perform multi-step \n reasoning and call tools\
\ reasoning and call tools for specific operations. In addition, there needs\ \ for specific operations. In addition, there needs to be general model-level\
\ to be general model-level \n safety checks as well as task-specific\ \ \n safety checks as well as task-specific safety checks that\
\ safety checks that are performed at a system level. \n\n We are\ \ are performed at a system level. \n\n We are defining the Llama\
\ defining the Llama Stack as a set of APIs and standards by synthesizing our\ \ Stack as a set of APIs and standards by synthesizing our learnings while \n\
\ learnings while \n working with Llama models. The APIs are divided\ \ working with Llama models. The APIs are divided into the llama-toolchain-api\
\ into the llama-toolchain-api and the llama-agentic-system-api. \n \ \ and the llama-agentic-system-api. \n These APIs provide a coherent\
\ These APIs provide a coherent way for model developers to fine tune and\ \ way for model developers to fine tune and serve Llama models, and agentic app\
\ serve Llama models, and agentic app \n developers to leverage\ \ \n developers to leverage all the capabilities of the Llama models\
\ all the capabilities of the Llama models seamlessly. We would like to work with\ \ seamlessly. We would like to work with the \n ecosystem to enhance\
\ the \n ecosystem to enhance and simplify the API. In addition,\ \ and simplify the API. In addition, we will be releasing a plug-in architecture\
\ we will be releasing a plug-in architecture to allow \n creating\ \ to allow \n creating distributions of the llama stack with different\
\ distributions of the llama stack with different implementations.\n\n\n \ \ implementations.\n\n\n This is the specification of the llama\
\ This is the specification of the llama stack that provides \n \ \ stack that provides \n a set of endpoints and their corresponding\
\ a set of endpoints and their corresponding interfaces that are tailored\ \ interfaces that are tailored to \n best leverage Llama Models.\
\ to \n best leverage Llama Models. The specification is still\ \ The specification is still in draft and subject to change."
\ in draft and subject to change."
title: '[DRAFT] Llama Stack Specification' title: '[DRAFT] Llama Stack Specification'
version: 0.0.1 version: 0.0.1
jsonSchemaDialect: https://json-schema.org/draft/2020-12/schema jsonSchemaDialect: https://json-schema.org/draft/2020-12/schema
@ -2053,13 +2052,13 @@ security:
servers: servers:
- url: http://any-hosted-llama-stack.com - url: http://any-hosted-llama-stack.com
tags: tags:
- name: PostTraining
- name: Inference
- name: MemoryBanks - name: MemoryBanks
- name: SyntheticDataGeneration
- name: RewardScoring
- name: AgenticSystem
- name: Datasets - name: Datasets
- name: AgenticSystem
- name: SyntheticDataGeneration
- name: PostTraining
- name: RewardScoring
- name: Inference
- description: <SchemaDefinition schemaRef="#/components/schemas/ShieldConfig" /> - description: <SchemaDefinition schemaRef="#/components/schemas/ShieldConfig" />
name: ShieldConfig name: ShieldConfig
- description: <SchemaDefinition schemaRef="#/components/schemas/AgenticSystemCreateRequest" - description: <SchemaDefinition schemaRef="#/components/schemas/AgenticSystemCreateRequest"