Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Use precomputed owl embeddings #973

Merged
merged 6 commits into from
Jan 24, 2025
Merged

Use precomputed owl embeddings #973

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
d9590cb
Use precomputed owl embeddings
probicheaux Jan 23, 2025
6e909fc
formatting
grzegorz-roboflow Jan 23, 2025
4a023cd
Merge branch 'main' into use-precomputed-owl-embeddings
grzegorz-roboflow Jan 23, 2025
ac748c5
Bump tokenizers dep (needed for owlvit)
probicheaux Jan 24, 2025
f710eab
less precise dep
probicheaux Jan 24, 2025
c191b5f
Merge branch 'main' into use-precomputed-owl-embeddings
grzegorz-roboflow Jan 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 24 additions & 3 deletions inference/models/owlv2/owlv2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,6 +46,8 @@
load_image_rgb,
)

CPU_IMAGE_EMBED_CACHE_SIZE = 10000

# TYPES
Hash = NewType("Hash", str)
PosNegKey = Literal["positive", "negative"]
Expand Down Expand Up @@ -319,7 +321,9 @@ def reset_cache(self):
# each entry should be on the order of 300*4KB, so 1000 is 400MB of CUDA memory
self.image_embed_cache = LimitedSizeDict(size_limit=OWLV2_IMAGE_CACHE_SIZE)
# no need for limit here, as we're only storing on CPU
self.cpu_image_embed_cache = dict()
self.cpu_image_embed_cache = LimitedSizeDict(
size_limit=CPU_IMAGE_EMBED_CACHE_SIZE
)
# each entry should be on the order of 10 bytes, so 1000 is 10KB
self.image_size_cache = LimitedSizeDict(size_limit=OWLV2_IMAGE_CACHE_SIZE)
# entry size will vary depending on the number of samples, but 10 should be safe
Expand Down Expand Up @@ -693,9 +697,24 @@ def serialize_training_data(
hf_id: str = f"google/{OWLV2_VERSION_ID}",
iou_threshold: float = 0.3,
save_dir: str = os.path.join(MODEL_CACHE_DIR, "owl-v2-serialized-data"),
previous_embeddings_file: str = None,
):
roboflow_id = hf_id.replace("google/", "owlv2/")
owlv2 = OwlV2(model_id=roboflow_id)
if previous_embeddings_file is not None:
if DEVICE == "cpu":
model_data = torch.load(previous_embeddings_file, map_location="cpu")
else:
model_data = torch.load(previous_embeddings_file)
grzegorz-roboflow marked this conversation as resolved.
Show resolved Hide resolved
class_names = model_data["class_names"]
train_data_dict = model_data["train_data_dict"]
huggingface_id = model_data["huggingface_id"]
roboflow_id = model_data["roboflow_id"]
# each model can have its own OwlV2 instance because we use a singleton
owlv2 = OwlV2(model_id=roboflow_id)
owlv2.cpu_image_embed_cache = model_data["image_embeds"]
else:
owlv2 = OwlV2(model_id=roboflow_id)

train_data_dict, image_embeds = owlv2.make_class_embeddings_dict(
training_data, iou_threshold, return_image_embeds=True
)
Expand Down Expand Up @@ -826,7 +845,9 @@ def draw_predictions(
def save_small_model_without_image_embeds(
self, save_dir: str = os.path.join(MODEL_CACHE_DIR, "owl-v2-serialized-data")
):
self.owlv2.cpu_image_embed_cache = dict()
self.owlv2.cpu_image_embed_cache = LimitedSizeDict(
size_limit=CPU_IMAGE_EMBED_CACHE_SIZE
)
return self.save_model(
self.huggingface_id,
self.roboflow_id,
Expand Down
2 changes: 1 addition & 1 deletion requirements/_requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,7 +34,7 @@ pandas>=2.0.0,<2.3.0
paho-mqtt~=1.6.1
pytest>=8.0.0,<9.0.0 # this is not a joke, sam2 requires this as the fork we are using is dependent on that, yet
# do not mark the dependency: https://github.com/SauravMaheshkar/samv2/blob/main/sam2/utils/download.py
tokenizers>=0.19.0,<=0.20.3
tokenizers>=0.19.0,<=0.21
slack-sdk~=3.33.4
twilio~=9.3.7
httpx>=0.25.1,<0.28.0 # must be pinned as bc in 0.28.0 is causing Anthropics to fail
Expand Down
Loading