mirror of
https://github.com/blakeblackshear/frigate.git
synced 2026-02-15 15:45:27 +03:00
Add device config for semantic search
This commit is contained in:
parent
1e6ee1a636
commit
f4eef74fdf
@ -276,7 +276,7 @@ class FrigateApp:
|
|||||||
def init_embeddings_client(self) -> None:
|
def init_embeddings_client(self) -> None:
|
||||||
if self.config.semantic_search.enabled:
|
if self.config.semantic_search.enabled:
|
||||||
# Create a client for other processes to use
|
# Create a client for other processes to use
|
||||||
self.embeddings = EmbeddingsContext(self.db)
|
self.embeddings = EmbeddingsContext(self.config, self.db)
|
||||||
|
|
||||||
def init_external_event_processor(self) -> None:
|
def init_external_event_processor(self) -> None:
|
||||||
self.external_event_processor = ExternalEventProcessor(self.config)
|
self.external_event_processor = ExternalEventProcessor(self.config)
|
||||||
|
|||||||
@ -12,3 +12,4 @@ class SemanticSearchConfig(FrigateBaseModel):
|
|||||||
reindex: Optional[bool] = Field(
|
reindex: Optional[bool] = Field(
|
||||||
default=False, title="Reindex all detections on startup."
|
default=False, title="Reindex all detections on startup."
|
||||||
)
|
)
|
||||||
|
device: str = Field(default="AUTO", title="Device Type")
|
||||||
|
|||||||
@ -55,7 +55,7 @@ def manage_embeddings(config: FrigateConfig) -> None:
|
|||||||
models = [Event]
|
models = [Event]
|
||||||
db.bind(models)
|
db.bind(models)
|
||||||
|
|
||||||
embeddings = Embeddings(db)
|
embeddings = Embeddings(config.semantic_search, db)
|
||||||
|
|
||||||
# Check if we need to re-index events
|
# Check if we need to re-index events
|
||||||
if config.semantic_search.reindex:
|
if config.semantic_search.reindex:
|
||||||
@ -70,8 +70,8 @@ def manage_embeddings(config: FrigateConfig) -> None:
|
|||||||
|
|
||||||
|
|
||||||
class EmbeddingsContext:
|
class EmbeddingsContext:
|
||||||
def __init__(self, db: SqliteVecQueueDatabase):
|
def __init__(self, config: FrigateConfig, db: SqliteVecQueueDatabase):
|
||||||
self.embeddings = Embeddings(db)
|
self.embeddings = Embeddings(config.semantic_search, db)
|
||||||
self.thumb_stats = ZScoreNormalization()
|
self.thumb_stats = ZScoreNormalization()
|
||||||
self.desc_stats = ZScoreNormalization()
|
self.desc_stats = ZScoreNormalization()
|
||||||
|
|
||||||
|
|||||||
@ -12,6 +12,7 @@ from PIL import Image
|
|||||||
from playhouse.shortcuts import model_to_dict
|
from playhouse.shortcuts import model_to_dict
|
||||||
|
|
||||||
from frigate.comms.inter_process import InterProcessRequestor
|
from frigate.comms.inter_process import InterProcessRequestor
|
||||||
|
from frigate.config.semantic_search import SemanticSearchConfig
|
||||||
from frigate.const import UPDATE_MODEL_STATE
|
from frigate.const import UPDATE_MODEL_STATE
|
||||||
from frigate.db.sqlitevecq import SqliteVecQueueDatabase
|
from frigate.db.sqlitevecq import SqliteVecQueueDatabase
|
||||||
from frigate.models import Event
|
from frigate.models import Event
|
||||||
@ -80,7 +81,10 @@ def deserialize(bytes_data: bytes) -> List[float]:
|
|||||||
class Embeddings:
|
class Embeddings:
|
||||||
"""SQLite-vec embeddings database."""
|
"""SQLite-vec embeddings database."""
|
||||||
|
|
||||||
def __init__(self, db: SqliteVecQueueDatabase) -> None:
|
def __init__(
|
||||||
|
self, config: SemanticSearchConfig, db: SqliteVecQueueDatabase
|
||||||
|
) -> None:
|
||||||
|
self.config = config
|
||||||
self.db = db
|
self.db = db
|
||||||
self.requestor = InterProcessRequestor()
|
self.requestor = InterProcessRequestor()
|
||||||
|
|
||||||
@ -118,7 +122,7 @@ class Embeddings:
|
|||||||
},
|
},
|
||||||
embedding_function=jina_text_embedding_function,
|
embedding_function=jina_text_embedding_function,
|
||||||
model_type="text",
|
model_type="text",
|
||||||
force_cpu=True,
|
device="CPU",
|
||||||
)
|
)
|
||||||
|
|
||||||
self.vision_embedding = GenericONNXEmbedding(
|
self.vision_embedding = GenericONNXEmbedding(
|
||||||
@ -130,6 +134,7 @@ class Embeddings:
|
|||||||
},
|
},
|
||||||
embedding_function=jina_vision_embedding_function,
|
embedding_function=jina_vision_embedding_function,
|
||||||
model_type="vision",
|
model_type="vision",
|
||||||
|
device=self.config.device,
|
||||||
)
|
)
|
||||||
|
|
||||||
def _create_tables(self):
|
def _create_tables(self):
|
||||||
|
|||||||
@ -42,7 +42,7 @@ class GenericONNXEmbedding:
|
|||||||
embedding_function: Callable[[List[np.ndarray]], np.ndarray],
|
embedding_function: Callable[[List[np.ndarray]], np.ndarray],
|
||||||
model_type: str,
|
model_type: str,
|
||||||
tokenizer_file: Optional[str] = None,
|
tokenizer_file: Optional[str] = None,
|
||||||
force_cpu: bool = False,
|
device: str = "AUTO",
|
||||||
):
|
):
|
||||||
self.model_name = model_name
|
self.model_name = model_name
|
||||||
self.model_file = model_file
|
self.model_file = model_file
|
||||||
@ -51,7 +51,7 @@ class GenericONNXEmbedding:
|
|||||||
self.embedding_function = embedding_function
|
self.embedding_function = embedding_function
|
||||||
self.model_type = model_type # 'text' or 'vision'
|
self.model_type = model_type # 'text' or 'vision'
|
||||||
self.providers, self.provider_options = get_ort_providers(
|
self.providers, self.provider_options = get_ort_providers(
|
||||||
force_cpu=force_cpu, requires_fp16=True
|
force_cpu=device == "CPU", requires_fp16=True, openvino_device=device
|
||||||
)
|
)
|
||||||
|
|
||||||
self.download_path = os.path.join(MODEL_CACHE_DIR, self.model_name)
|
self.download_path = os.path.join(MODEL_CACHE_DIR, self.model_name)
|
||||||
|
|||||||
@ -42,7 +42,7 @@ class EmbeddingMaintainer(threading.Thread):
|
|||||||
threading.Thread.__init__(self)
|
threading.Thread.__init__(self)
|
||||||
self.name = "embeddings_maintainer"
|
self.name = "embeddings_maintainer"
|
||||||
self.config = config
|
self.config = config
|
||||||
self.embeddings = Embeddings(db)
|
self.embeddings = Embeddings(config.semantic_search, db)
|
||||||
self.event_subscriber = EventUpdateSubscriber()
|
self.event_subscriber = EventUpdateSubscriber()
|
||||||
self.event_end_subscriber = EventEndSubscriber()
|
self.event_end_subscriber = EventEndSubscriber()
|
||||||
self.event_metadata_subscriber = EventMetadataSubscriber(
|
self.event_metadata_subscriber = EventMetadataSubscriber(
|
||||||
|
|||||||
@ -36,7 +36,7 @@ class EventCleanup(threading.Thread):
|
|||||||
self.camera_labels: dict[str, dict[str, any]] = {}
|
self.camera_labels: dict[str, dict[str, any]] = {}
|
||||||
|
|
||||||
if self.config.semantic_search.enabled:
|
if self.config.semantic_search.enabled:
|
||||||
self.embeddings = Embeddings(self.db)
|
self.embeddings = Embeddings(self.config.semantic_search, self.db)
|
||||||
|
|
||||||
def get_removed_camera_labels(self) -> list[Event]:
|
def get_removed_camera_labels(self) -> list[Event]:
|
||||||
"""Get a list of distinct labels for removed cameras."""
|
"""Get a list of distinct labels for removed cameras."""
|
||||||
|
|||||||
Loading…
Reference in New Issue
Block a user