Docker Slim Image

.dockerignore

@@ -3,11 +3,15 @@
 __pycache__/
 docs/
 
+.vscode
+.git
+.mypy_cache
+.ruff_cache
+.pytype
 .coverage
 .coverage.*
 .coverage/
 coverage.xml
 .readthedocs.yml
-*.toml
 
 !README.md

Dockerfile.slim

@@ -0,0 +1,71 @@
+# Copyright (c) MONAI Consortium
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#     http://www.apache.org/licenses/LICENSE-2.0
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# To build with a different base image
+# please run `docker build` using the `--build-arg IMAGE=...` flag.
+ARG IMAGE=debian:12-slim
+
+FROM ${IMAGE} AS build
+
+ARG TORCH_CUDA_ARCH_LIST="7.5 8.0 8.6 8.9 9.0+PTX"
+
+ENV DEBIAN_FRONTEND=noninteractive
+ENV APT_INSTALL="apt install -y --no-install-recommends"
+
+RUN apt update && apt upgrade -y && \
+    ${APT_INSTALL} ca-certificates python3-pip python-is-python3 git wget libopenslide0 unzip python3-dev && \
+    wget https://developer.download.nvidia.com/compute/cuda/repos/debian12/x86_64/cuda-keyring_1.1-1_all.deb && \
+    dpkg -i cuda-keyring_1.1-1_all.deb && \
+    apt update && \
+    ${APT_INSTALL} cuda-toolkit-12 && \
+    apt clean && \
+    rm -rf cuda-keyring_1.1-1_all.deb /usr/lib/python*/EXTERNALLY-MANAGED /var/lib/apt/lists/*
+
+# TODO: remark for issue [revise the dockerfile](https://github.com/zarr-developers/numcodecs/issues/431)
+RUN if [[ $(uname -m) =~ "aarch64" ]]; then \
+      CFLAGS="-O3" DISABLE_NUMCODECS_SSE2=true DISABLE_NUMCODECS_AVX2=true python -m pip install numcodecs; \
+    fi
+
+# NGC Client
+WORKDIR /opt/tools
+ARG NGC_CLI_URI="https://ngc.nvidia.com/downloads/ngccli_linux.zip"
+RUN wget -q ${NGC_CLI_URI} && unzip ngccli_linux.zip && chmod u+x ngc-cli/ngc && \
+    find ngc-cli/ -type f -exec md5sum {} + | LC_ALL=C sort | md5sum -c ngc-cli.md5 && \
+    rm -rf ngccli_linux.zip ngc-cli.md5
+
+WORKDIR /opt/monai
+
+# copy relevant parts of repo
+COPY requirements.txt requirements-min.txt requirements-dev.txt versioneer.py setup.py setup.cfg pyproject.toml ./
+COPY LICENSE CHANGELOG.md CODE_OF_CONDUCT.md CONTRIBUTING.md README.md MANIFEST.in runtests.sh ./
+COPY tests ./tests
+COPY monai ./monai
+
+# install full deps
+RUN python -m pip install --upgrade --no-cache-dir pip && \
+    python -m pip install --no-cache-dir -r requirements-dev.txt
+
+# compile ext and remove temp files
+RUN export CUDA_HOME=/usr/local/cuda FORCE_CUDA=1 USE_COMPILED=1 BUILD_MONAI=1 && \
+    python setup.py develop && \
+    rm -rf build __pycache__ monai.egg-info
+
+# flatten all layers down to one
+FROM ${IMAGE}
+LABEL maintainer="[email protected]"
+
+WORKDIR /opt/monai
+COPY --from=build / /
+
+ENV PATH=${PATH}:/opt/tools:/opt/tools/ngc-cli
+ENV POLYGRAPHY_AUTOINSTALL_DEPS=1
+ENV CUDA_HOME=/usr/local/cuda
+ENV BUILD_MONAI=1

monai/apps/vista3d/inferer.py

@@ -86,13 +86,13 @@ def point_based_window_inferer(
             for j in range(len(ly_)):
                 for k in range(len(lz_)):
                     lx, rx, ly, ry, lz, rz = (lx_[i], rx_[i], ly_[j], ry_[j], lz_[k], rz_[k])
-                    unravel_slice = [
+                    unravel_slice = (
                         slice(None),
                         slice(None),
                         slice(int(lx), int(rx)),
                         slice(int(ly), int(ry)),
                         slice(int(lz), int(rz)),
-                    ]
+                    )
                     batch_image = image[unravel_slice]
                     output = predictor(
                         batch_image,

monai/networks/nets/vista3d.py

@@ -243,14 +243,10 @@ def connected_components_combine(
         _logits = logits[mapping_index]
         inside = []
         for i in range(_logits.shape[0]):
-            inside.append(
-                np.any(
-                    [
-                        _logits[i, 0, p[0], p[1], p[2]].item() > 0
-                        for p in point_coords[i].cpu().numpy().round().astype(int)
-                    ]
-                )
-            )
+            p_coord = point_coords[i].cpu().numpy().round().astype(int)
+            inside_p = [_logits[i, 0, p[0], p[1], p[2]].item() > 0 for p in p_coord]
+            inside.append(int(np.any(inside_p)))  # convert to int to avoid typing problems with Numpy
+
         inside_tensor = torch.tensor(inside).to(logits.device)
         nan_mask = torch.isnan(_logits)
         # _logits are converted to binary [B1, 1, H, W, D]

monai/networks/utils.py

@@ -713,13 +713,13 @@ def convert_to_onnx(
         torch_versioned_kwargs = {}
         if use_trace:
             # let torch.onnx.export to trace the model.
-            mode_to_export = model
+            model_to_export = model
             torch_versioned_kwargs = kwargs
             if "dynamo" in kwargs and kwargs["dynamo"] and verify:
                 torch_versioned_kwargs["verify"] = verify
                 verify = False
         else:
-            mode_to_export = torch.jit.script(model, **kwargs)
+            model_to_export = torch.jit.script(model, **kwargs)
 
         if torch.is_tensor(inputs) or isinstance(inputs, dict):
             onnx_inputs = (inputs,)
@@ -733,7 +733,7 @@ def convert_to_onnx(
             f = filename
         print(f"torch_versioned_kwargs={torch_versioned_kwargs}")
         torch.onnx.export(
-            mode_to_export,
+            model_to_export,
             onnx_inputs,
             f=f,
             input_names=input_names,

requirements-dev.txt

@@ -1,6 +1,6 @@
 # Full requirements for developments
 -r requirements-min.txt
-pytorch-ignite==0.4.11
+pytorch-ignite
 gdown>=4.7.3
 scipy>=1.12.0; python_version >= '3.9'
 itk>=5.2
@@ -51,8 +51,9 @@ h5py
 nni==2.10.1; platform_system == "Linux" and "arm" not in platform_machine and "aarch" not in platform_machine
 optuna
 git+https://github.com/Project-MONAI/MetricsReloaded@monai-support#egg=MetricsReloaded
-onnx>=1.13.0, <1.19.1
-onnxruntime; python_version <= '3.10'
+onnx>=1.13.0, <1.19.0
+onnxruntime
+onnxscript
 typeguard<3  # https://github.com/microsoft/nni/issues/5457
 filelock<3.12.0  # https://github.com/microsoft/nni/issues/5523
 zarr

requirements.txt

@@ -1,3 +1,3 @@
-torch>=2.4.1; platform_system != "Windows"
-torch>=2.4.1, !=2.7.0; platform_system == "Windows"
+torch>=2.4.1, <2.9; platform_system != "Windows"
+torch>=2.4.1, <2.9, !=2.7.0; platform_system == "Windows"
 numpy>=1.24,<3.0

tests/bundle/test_bundle_download.py

@@ -15,7 +15,7 @@
 import os
 import tempfile
 import unittest
-from unittest.case import skipUnless
+from unittest.case import skipIf, skipUnless
 from unittest.mock import patch
 
 import numpy as np
@@ -219,6 +219,7 @@ def test_monaihosting_url_download_bundle(self, bundle_files, bundle_name, url):
 
     @parameterized.expand([TEST_CASE_5])
     @skip_if_quick
+    @skipIf(os.getenv("NGC_API_KEY", None) is None, "NGC API key required for this test")
     def test_ngc_private_source_download_bundle(self, bundle_files, bundle_name, _url):
         with skip_if_downloading_fails():
             # download a single file from url, also use `args_file`

tests/data/meta_tensor/test_meta_tensor.py

@@ -245,7 +245,7 @@ def test_pickling(self):
         with tempfile.TemporaryDirectory() as tmp_dir:
             fname = os.path.join(tmp_dir, "im.pt")
             torch.save(m, fname)
-            m2 = torch.load(fname, weights_only=True)
+            m2 = torch.load(fname, weights_only=False)
         self.check(m2, m, ids=False)
 
     @skip_if_no_cuda

tests/losses/test_multi_scale.py

@@ -55,7 +55,7 @@ class TestMultiScale(unittest.TestCase):
     @parameterized.expand(TEST_CASES)
     def test_shape(self, input_param, input_data, expected_val):
         result = MultiScaleLoss(**input_param).forward(**input_data)
-        np.testing.assert_allclose(result.detach().cpu().numpy(), expected_val, rtol=1e-5)
+        np.testing.assert_allclose(result.detach().cpu().numpy(), expected_val, rtol=1e-4)
 
     @parameterized.expand(
         [

tests/networks/test_convert_to_onnx.py

@@ -22,10 +22,14 @@
 from monai.networks.nets import SegResNet, UNet
 from tests.test_utils import SkipIfNoModule, optional_import, skip_if_quick
 
-if torch.cuda.is_available():
-    TORCH_DEVICE_OPTIONS = ["cpu", "cuda"]
-else:
-    TORCH_DEVICE_OPTIONS = ["cpu"]
+onnx, _ = optional_import("onnx")
+
+TORCH_DEVICE_OPTIONS = ["cpu"]
+
+# FIXME: CUDA seems to produce different model outputs during testing vs. ONNX outputs, use CPU only for now
+# if torch.cuda.is_available():
+#     TORCH_DEVICE_OPTIONS.append("cuda")
+
 TESTS = list(itertools.product(TORCH_DEVICE_OPTIONS, [True, False], [True, False]))
 TESTS_ORT = list(itertools.product(TORCH_DEVICE_OPTIONS, [True]))
 
@@ -35,38 +39,38 @@
 else:
     rtol, atol = 1e-3, 1e-4
 
-onnx, _ = optional_import("onnx")
-
 
 @SkipIfNoModule("onnx")
 @skip_if_quick
 class TestConvertToOnnx(unittest.TestCase):
     @parameterized.expand(TESTS)
     def test_unet(self, device, use_trace, use_ort):
+        """Test converting UNet to ONNX."""
         if use_ort:
             _, has_onnxruntime = optional_import("onnxruntime")
             if not has_onnxruntime:
                 self.skipTest("onnxruntime is not installed probably due to python version >= 3.11.")
         model = UNet(
             spatial_dims=2, in_channels=1, out_channels=3, channels=(16, 32, 64), strides=(2, 2), num_res_units=0
         )
-        if use_trace:
-            onnx_model = convert_to_onnx(
-                model=model,
-                inputs=[torch.randn((16, 1, 32, 32), requires_grad=False)],
-                input_names=["x"],
-                output_names=["y"],
-                verify=True,
-                device=device,
-                use_ort=use_ort,
-                use_trace=use_trace,
-                rtol=rtol,
-                atol=atol,
-            )
-            self.assertTrue(isinstance(onnx_model, onnx.ModelProto))
+
+        onnx_model = convert_to_onnx(
+            model=model,
+            inputs=[torch.randn((16, 1, 32, 32), requires_grad=False)],
+            input_names=["x"],
+            output_names=["y"],
+            verify=True,
+            device=device,
+            use_ort=use_ort,
+            use_trace=use_trace,
+            rtol=rtol,
+            atol=atol,
+        )
+        self.assertTrue(isinstance(onnx_model, onnx.ModelProto))
 
     @parameterized.expand(TESTS_ORT)
     def test_seg_res_net(self, device, use_ort):
+        """Test converting SetResNet to ONNX."""
         if use_ort:
             _, has_onnxruntime = optional_import("onnxruntime")
             if not has_onnxruntime: