administrator/frigate
1
0
Fork 0
mirror of https://github.com/blakeblackshear/frigate.git synced 2026-01-31 16:25:26 +03:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

add locks to jina v1 embeddings

Browse Source 
protect tokenizer and feature extractor in jina_v1_embedding with per-instance thread lock to avoid the "Already borrowed" RuntimeError during concurrent tokenization
This commit is contained in:
Josh Hawkins 2026-01-14 16:26:25 -06:00
parent 7bc2ef731b
commit 2c9f7a5275
1 changed files with 21 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
frigate/embeddings/onnx/jina_v1_embedding.py
View File

@ -2,6 +2,7 @@
import logging import logging
import os import os
import threading
import warnings import warnings
from transformers import AutoFeatureExtractor, AutoTokenizer from transformers import AutoFeatureExtractor, AutoTokenizer
@ -54,6 +55,7 @@ class JinaV1TextEmbedding(BaseEmbedding):
self.tokenizer = None self.tokenizer = None
self.feature_extractor = None self.feature_extractor = None
self.runner = None self.runner = None
self._lock = threading.Lock()
files_names = list(self.download_urls.keys()) + [self.tokenizer_file] files_names = list(self.download_urls.keys()) + [self.tokenizer_file]
if not all( if not all(
@ -134,6 +136,7 @@ class JinaV1TextEmbedding(BaseEmbedding):
) )
def _preprocess_inputs(self, raw_inputs): def _preprocess_inputs(self, raw_inputs):
with self._lock:
max_length = max(len(self.tokenizer.encode(text)) for text in raw_inputs) max_length = max(len(self.tokenizer.encode(text)) for text in raw_inputs)
return [ return [
self.tokenizer( self.tokenizer(
@ -174,6 +177,7 @@ class JinaV1ImageEmbedding(BaseEmbedding):
self.download_path = os.path.join(MODEL_CACHE_DIR, self.model_name) self.download_path = os.path.join(MODEL_CACHE_DIR, self.model_name)
self.feature_extractor = None self.feature_extractor = None
self.runner: BaseModelRunner | None = None self.runner: BaseModelRunner | None = None
self._lock = threading.Lock()
files_names = list(self.download_urls.keys()) files_names = list(self.download_urls.keys())
if not all( if not all(
os.path.exists(os.path.join(self.download_path, n)) for n in files_names os.path.exists(os.path.join(self.download_path, n)) for n in files_names
@ -216,6 +220,7 @@ class JinaV1ImageEmbedding(BaseEmbedding):
) )
def _preprocess_inputs(self, raw_inputs): def _preprocess_inputs(self, raw_inputs):
with self._lock:
processed_images = [self._process_image(img) for img in raw_inputs] processed_images = [self._process_image(img) for img in raw_inputs]
return [ return [
self.feature_extractor(images=image, return_tensors="np") self.feature_extractor(images=image, return_tensors="np")