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

Jump to bottom

cleanup: Remove manual offload from Depth Anything Processor #5812

Merged
merged 2 commits into from
Mar 1, 2024
Merged

cleanup: Remove manual offload from Depth Anything Processor #5812

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6f77477
cleanup: remove manual offload from depth anything
blessedcoolant Feb 27, 2024
41b77cd
fix: minor fixes to types in the DA Detector
blessedcoolant Feb 27, 2024
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
5 changes: 2 additions & 3 deletions invokeai/app/invocations/controlnet_image_processors.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -576,7 +576,7 @@ def run_processor(self, image: Image.Image):
title="Depth Anything Processor",
tags=["controlnet", "depth", "depth anything"],
category="controlnet",
version="1.0.0",
version="1.0.1",
)
class DepthAnythingImageProcessorInvocation(ImageProcessorInvocation):
"""Generates a depth map based on the Depth Anything algorithm"""
Expand All @@ -585,13 +585,12 @@ class DepthAnythingImageProcessorInvocation(ImageProcessorInvocation):
default="small", description="The size of the depth model to use"
)
resolution: int = InputField(default=512, ge=64, multiple_of=64, description=FieldDescriptions.image_res)
offload: bool = InputField(default=False)

def run_processor(self, image: Image.Image):
depth_anything_detector = DepthAnythingDetector()
depth_anything_detector.load_model(model_size=self.model_size)

processed_image = depth_anything_detector(image=image, resolution=self.resolution, offload=self.offload)
processed_image = depth_anything_detector(image=image, resolution=self.resolution)
return processed_image


Expand Down
26 changes: 14 additions & 12 deletions invokeai/backend/image_util/depth_anything/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,8 @@

config = InvokeAIAppConfig.get_config()

DEPTH_ANYTHING_MODEL_SIZES = Literal["large", "base", "small"]

DEPTH_ANYTHING_MODELS = {
"large": {
"url": "https://huggingface.co/spaces/LiheYoung/Depth-Anything/resolve/main/checkpoints/depth_anything_vitl14.pth?download=true",
Expand Down Expand Up @@ -53,9 +55,9 @@
class DepthAnythingDetector:
def __init__(self) -> None:
self.model = None
self.model_size: Union[Literal["large", "base", "small"], None] = None
self.model_size: Union[DEPTH_ANYTHING_MODEL_SIZES, None] = None

def load_model(self, model_size=Literal["large", "base", "small"]):
def load_model(self, model_size: DEPTH_ANYTHING_MODEL_SIZES = "small"):
DEPTH_ANYTHING_MODEL_PATH = pathlib.Path(config.models_path / DEPTH_ANYTHING_MODELS[model_size]["local"])
if not DEPTH_ANYTHING_MODEL_PATH.exists():
download_with_progress_bar(DEPTH_ANYTHING_MODELS[model_size]["url"], DEPTH_ANYTHING_MODEL_PATH)
Expand Down Expand Up @@ -84,16 +86,19 @@ def to(self, device):
self.model.to(device)
return self

def __call__(self, image, resolution=512, offload=False):
image = np.array(image, dtype=np.uint8)
image = image[:, :, ::-1] / 255.0
def __call__(self, image: Image.Image, resolution: int = 512):
if self.model is None:
raise Exception("Depth Anything Model not loaded")

np_image = np.array(image, dtype=np.uint8)
np_image = np_image[:, :, ::-1] / 255.0

image_height, image_width = image.shape[:2]
image = transform({"image": image})["image"]
image = torch.from_numpy(image).unsqueeze(0).to(choose_torch_device())
image_height, image_width = np_image.shape[:2]
np_image = transform({"image": image})["image"]
tensor_image = torch.from_numpy(np_image).unsqueeze(0).to(choose_torch_device())

with torch.no_grad():
depth = self.model(image)
depth = self.model(tensor_image)
depth = F.interpolate(depth[None], (image_height, image_width), mode="bilinear", align_corners=False)[0, 0]
depth = (depth - depth.min()) / (depth.max() - depth.min()) * 255.0

Expand All @@ -103,7 +108,4 @@ def __call__(self, image, resolution=512, offload=False):
new_height = int(image_height * (resolution / image_width))
depth_map = depth_map.resize((resolution, new_height))

if offload:
del self.model

return depth_map
Loading