nan
2ff7b04f56
- Introduced a new memory management policy for ONNX Runtime to release temporary GPU memory after inference completion. - Established a shared inference entry point for all ONNX models, differentiating between long-lived sessions and temporary memory arenas. - Updated CUDA sessions to utilize a shrinkable memory arena, reclaiming GPU memory post-inference. - Ensured consistent memory release semantics across BEN2, DA3, MoGe-2, and Upscale models while reusing existing sessions. - Added lifecycle tests to validate temporary memory release and session reuse. - Modified model adapters to integrate the new memory management features, ensuring backward compatibility with existing inference behavior. - Enhanced unit tests to cover new memory release scenarios and verify correct behavior across different execution providers.
93 lines
2.8 KiB
Python
93 lines
2.8 KiB
Python
import sys
|
|
import tempfile
|
|
import unittest
|
|
from pathlib import Path
|
|
from types import SimpleNamespace
|
|
from unittest.mock import patch
|
|
|
|
import numpy as np
|
|
from PIL import Image
|
|
|
|
|
|
SOURCE_DIR = (
|
|
Path(__file__).resolve().parents[1]
|
|
/ "src"
|
|
/ "anyimage"
|
|
/ "server"
|
|
/ "models"
|
|
)
|
|
sys.path.insert(0, str(SOURCE_DIR))
|
|
|
|
import onnx_da3
|
|
|
|
|
|
class FakeSession:
|
|
def get_providers(self):
|
|
return ["CPUExecutionProvider"]
|
|
|
|
def get_inputs(self):
|
|
return [SimpleNamespace(name="images"), SimpleNamespace(name="strategy")]
|
|
|
|
def run(self, output_names, feeds):
|
|
self.feeds = feeds
|
|
_, views, _, height, width = feeds["images"].shape
|
|
depth = np.ones((1, views, height, width), dtype=np.float32)
|
|
confidence = np.full_like(depth, 0.75)
|
|
pose = np.zeros((1, views, 9), dtype=np.float32)
|
|
pose[..., 6] = 1.0
|
|
pose[..., 7:] = np.pi / 2.0
|
|
return depth, confidence, pose
|
|
|
|
|
|
class Da3OnnxTest(unittest.TestCase):
|
|
def test_infer_forwards_memory_release_boundary(self):
|
|
with tempfile.TemporaryDirectory() as directory:
|
|
image_path = Path(directory) / "input.png"
|
|
Image.new("RGB", (80, 60)).save(image_path)
|
|
session = FakeSession()
|
|
with patch.object(
|
|
onnx_da3,
|
|
"run_session",
|
|
side_effect=lambda model, names, feeds, **_options: model.run(
|
|
names, feeds
|
|
),
|
|
) as run:
|
|
onnx_da3.infer(
|
|
session,
|
|
[image_path],
|
|
process_res=56,
|
|
release_memory=False,
|
|
)
|
|
|
|
self.assertFalse(run.call_args.kwargs["release_memory"])
|
|
|
|
def test_target_size_uses_patch_multiple(self):
|
|
self.assertEqual(
|
|
onnx_da3.target_size((800, 600), 504, "upper_bound_resize"),
|
|
(504, 378),
|
|
)
|
|
|
|
def test_infer_returns_camera_and_image_arrays(self):
|
|
with tempfile.TemporaryDirectory() as directory:
|
|
image_path = Path(directory) / "input.png"
|
|
Image.new("RGB", (80, 60), (10, 20, 30)).save(image_path)
|
|
session = FakeSession()
|
|
|
|
result = onnx_da3.infer(
|
|
session,
|
|
[image_path],
|
|
process_res=56,
|
|
ref_view_strategy="middle",
|
|
)
|
|
|
|
self.assertEqual(result.depth.shape, (1, 42, 56))
|
|
self.assertEqual(result.conf.shape, result.depth.shape)
|
|
self.assertEqual(result.processed_images.shape, (1, 42, 56, 3))
|
|
self.assertEqual(result.extrinsics.shape, (1, 3, 4))
|
|
self.assertEqual(result.intrinsics.shape, (1, 3, 3))
|
|
self.assertEqual(int(session.feeds["strategy"][0]), 1)
|
|
np.testing.assert_allclose(result.extrinsics[0, :3, :3], np.eye(3))
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|