immich-app/immich
1
0
Fork 0
mirror of https://github.com/immich-app/immich.git synced 2025-12-24 01:11:32 +03:00
Code Issues Packages Projects Releases 18 Wiki Activity

feat(ml): composable ml (#9973)

Browse Source 
* modularize model classes

* various fixes

* expose port

* change response

* round coordinates

* simplify preload

* update server

* simplify interface

simplify

* update tests

* composable endpoint

* cleanup

fixes

remove unnecessary interface

support text input, cleanup

* ew camelcase

* update server

server fixes

fix typing

* ml fixes

update locustfile

fixes

* cleaner response

* better repo response

* update tests

formatting and typing

rename

* undo compose change

* linting

fix type

actually fix typing

* stricter typing

fix detection-only response

no need for defaultdict

* update spec file

update api

linting

* update e2e

* unnecessary dimension

* remove commented code

* remove duplicate code

* remove unused imports

* add batch dim
This commit is contained in:
Mert
2024-06-06 23:09:47 -04:00
committed by GitHub
parent 7a46f80ddc
commit 2b1b43a7e4
39 changed files with 982 additions and 999 deletions
Download Patch File Download Diff File Expand all files Collapse all files

50
machine-learning/app/models/__init__.py
View File

@@ -1,24 +1,40 @@
from typing import Any
from app.schemas import ModelType
from app.models.base import InferenceModel
from app.models.clip.textual import MClipTextualEncoder, OpenClipTextualEncoder
from app.models.clip.visual import OpenClipVisualEncoder
from app.schemas import ModelSource, ModelTask, ModelType
from .base import InferenceModel
from .clip import MCLIPEncoder, OpenCLIPEncoder
from .constants import is_insightface, is_mclip, is_openclip
from .facial_recognition import FaceRecognizer
from .constants import get_model_source
from .facial_recognition.detection import FaceDetector
from .facial_recognition.recognition import FaceRecognizer
def from_model_type(model_type: ModelType, model_name: str, **model_kwargs: Any) -> InferenceModel:
match model_type:
case ModelType.CLIP:
if is_openclip(model_name):
return OpenCLIPEncoder(model_name, **model_kwargs)
elif is_mclip(model_name):
return MCLIPEncoder(model_name, **model_kwargs)
case ModelType.FACIAL_RECOGNITION:
if is_insightface(model_name):
return FaceRecognizer(model_name, **model_kwargs)
def get_model_class(model_name: str, model_type: ModelType, model_task: ModelTask) -> type[InferenceModel]:
source = get_model_source(model_name)
match source, model_type, model_task:
case ModelSource.OPENCLIP | ModelSource.MCLIP, ModelType.VISUAL, ModelTask.SEARCH:
return OpenClipVisualEncoder
case ModelSource.OPENCLIP, ModelType.TEXTUAL, ModelTask.SEARCH:
return OpenClipTextualEncoder
case ModelSource.MCLIP, ModelType.TEXTUAL, ModelTask.SEARCH:
return MClipTextualEncoder
case ModelSource.INSIGHTFACE, ModelType.DETECTION, ModelTask.FACIAL_RECOGNITION:
return FaceDetector
case ModelSource.INSIGHTFACE, ModelType.RECOGNITION, ModelTask.FACIAL_RECOGNITION:
return FaceRecognizer
case _:
raise ValueError(f"Unknown model type {model_type}")
raise ValueError(f"Unknown model combination: {source}, {model_type}, {model_task}")
raise ValueError(f"Unknown {model_type} model {model_name}")
def from_model_type(model_name: str, model_type: ModelType, model_task: ModelTask, **kwargs: Any) -> InferenceModel:
return get_model_class(model_name, model_type, model_task)(model_name, **kwargs)
def get_model_deps(model_name: str, model_type: ModelType, model_task: ModelTask) -> list[tuple[ModelType, ModelTask]]:
return get_model_class(model_name, model_type, model_task).depends