[Evals API][4/n] evals with generation meta-reference impl (#303)

* wip

* dataset validation

* test_scoring

* cleanup

* clean up test

* comments

* error checking

* dataset client

* test client:

* datasetio client

* clean up

* basic scoring function works

* scorer wip

* equality scorer

* score batch impl

* score batch

* update scoring test

* refactor

* validate scorer input

* address comments

* evals with generation

* add all rows scores to ScoringResult

* minor typing

* bugfix

* scoring function def rename

* rebase name

* refactor

* address comments

* Update iOS inference instructions for new quantization

* Small updates to quantization config

* Fix score threshold in faiss

* Bump version to 0.0.45

* Handle both ipv6 and ipv4 interfaces together

* update manifest for build templates

* Update getting_started.md

* chatcompletion & completion input type validation

* inclusion->subsetof

* error checking

* scoring_function -> scoring_fn rename, scorer -> scoring_fn rename

* address comments

* [Evals API][5/n] fixes to generate openapi spec (#323)

* generate openapi

* typing comment, dataset -> dataset_id

* remove custom type

* sample eval run.yaml

---------

Co-authored-by: Dalton Flanagan <6599399+dltn@users.noreply.github.com>
Co-authored-by: Ashwin Bharambe <ashwin.bharambe@gmail.com>
This commit is contained in:
Xi Yan 2024-10-25 13:12:39 -07:00 committed by GitHub
parent 426d821e7f
commit abdf7cddf3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
31 changed files with 3371 additions and 1296 deletions

View file

@ -3,6 +3,8 @@
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from enum import Enum
from llama_models.schema_utils import json_schema_type
from pydantic import BaseModel
@ -10,3 +12,9 @@ from pydantic import BaseModel
@json_schema_type
class Job(BaseModel):
job_id: str
@json_schema_type
class JobStatus(Enum):
completed = "completed"
in_progress = "in_progress"

View file

@ -4,7 +4,7 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from typing import Dict, List, Literal, Union
from typing import Literal, Union
from pydantic import BaseModel, Field
from typing_extensions import Annotated
@ -24,12 +24,10 @@ class BooleanType(BaseModel):
class ArrayType(BaseModel):
type: Literal["array"] = "array"
items: "ParamType"
class ObjectType(BaseModel):
type: Literal["object"] = "object"
properties: Dict[str, "ParamType"] = Field(default_factory=dict)
class JsonType(BaseModel):
@ -38,12 +36,21 @@ class JsonType(BaseModel):
class UnionType(BaseModel):
type: Literal["union"] = "union"
options: List["ParamType"] = Field(default_factory=list)
class CustomType(BaseModel):
type: Literal["custom"] = "custom"
validator_class: str
class ChatCompletionInputType(BaseModel):
# expects List[Message] for messages
type: Literal["chat_completion_input"] = "chat_completion_input"
class CompletionInputType(BaseModel):
# expects InterleavedTextMedia for content
type: Literal["completion_input"] = "completion_input"
class AgentTurnInputType(BaseModel):
# expects List[Message] for messages (may also include attachments?)
type: Literal["agent_turn_input"] = "agent_turn_input"
ParamType = Annotated[
@ -55,11 +62,22 @@ ParamType = Annotated[
ObjectType,
JsonType,
UnionType,
CustomType,
ChatCompletionInputType,
CompletionInputType,
AgentTurnInputType,
],
Field(discriminator="type"),
]
ArrayType.model_rebuild()
ObjectType.model_rebuild()
UnionType.model_rebuild()
# TODO: recursive definition of ParamType in these containers
# will cause infinite recursion in OpenAPI generation script
# since we are going with ChatCompletionInputType and CompletionInputType
# we don't need to worry about ArrayType/ObjectType/UnionType for now
# ArrayType.model_rebuild()
# ObjectType.model_rebuild()
# UnionType.model_rebuild()
# class CustomType(BaseModel):
# type: Literal["custom"] = "custom"
# validator_class: str