Skip to content

[Enhancement] Load openvino model from memory #2027

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

Merged
merged 3 commits into from
Apr 28, 2023
Merged

[Enhancement] Load openvino model from memory #2027

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
2f491ec
fix cpu docker (different openvino ver for python & sdk)
irexyc Apr 25, 2023
25a630a
load openvino model from memory
irexyc Apr 25, 2023
4be2f37
fix lint
irexyc Apr 25, 2023
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
24 changes: 6 additions & 18 deletions csrc/mmdeploy/net/openvino/openvino_net.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,30 +77,18 @@ Result<void> OpenVINONet::Init(const Value& args) {
auto model = context["model"].get<Model>();
OUTCOME_TRY(auto config, model.GetModelConfig(name));

// TODO: read network with stream
// save xml and bin to temp file
auto tmp_dir = fs::temp_directory_path();
std::string tmp_xml = (tmp_dir / fs::path("tmp.xml")).string();
std::string tmp_bin = (tmp_dir / fs::path("tmp.bin")).string();
OUTCOME_TRY(auto raw_xml, model.ReadFile(config.net));
OUTCOME_TRY(auto raw_bin, model.ReadFile(config.weights));

try {
std::ofstream xml_out(tmp_xml, std::ios::binary);
xml_out << raw_xml;
xml_out.close();
std::ofstream bin_out(tmp_bin, std::ios::binary);
bin_out << raw_bin;
bin_out.close();
} catch (const std::exception& e) {
MMDEPLOY_ERROR("unhandled exception when creating tmp xml/bin: {}", e.what());
return Status(eFail);
}
auto ov_tensor = InferenceEngine::TensorDesc(InferenceEngine::Precision::U8, {raw_bin.size()},
InferenceEngine::Layout::C);
auto ov_blob = InferenceEngine::make_shared_blob<uint8_t>(ov_tensor);
ov_blob->allocate();
memcpy(ov_blob->buffer(), raw_bin.data(), ov_blob->byteSize());

try {
// create cnnnetwork
core_ = InferenceEngine::Core();
network_ = core_.ReadNetwork(tmp_xml, tmp_bin);
network_ = core_.ReadNetwork(raw_xml, std::move(ov_blob));

// set input tensor
InferenceEngine::InputsDataMap input_info = network_.getInputsInfo();
Expand Down
10 changes: 5 additions & 5 deletions docker/CPU/Dockerfile
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
FROM openvino/ubuntu18_dev:2021.4.2
FROM openvino/ubuntu20_dev:2022.3.0
ARG PYTHON_VERSION=3.8
ARG TORCH_VERSION=1.10.0
ARG TORCHVISION_VERSION=0.11.0
Expand Down Expand Up @@ -57,7 +57,7 @@ RUN if [ ${USE_SRC_INSIDE} == true ] ; \
RUN /opt/conda/bin/pip install torch==${TORCH_VERSION}+cpu torchvision==${TORCHVISION_VERSION}+cpu -f https://download.pytorch.org/whl/cpu/torch_stable.html \
&& /opt/conda/bin/pip install --no-cache-dir openmim

RUN /opt/conda/bin/mim install --no-cache-dir "mmcv"${MMCV_VERSION} onnxruntime==${ONNXRUNTIME_VERSION} openvino-dev mmengine${MMENGINE_VERSION}
RUN /opt/conda/bin/mim install --no-cache-dir "mmcv"${MMCV_VERSION} onnxruntime==${ONNXRUNTIME_VERSION} openvino-dev==2022.3.0 mmengine${MMENGINE_VERSION}

ENV PATH /opt/conda/bin:$PATH
WORKDIR /root/workspace
Expand Down Expand Up @@ -100,14 +100,14 @@ RUN git clone -b main https://github.com/open-mmlab/mmdeploy.git &&\
/opt/conda/bin/mim install -e .

### build SDK
ENV LD_LIBRARY_PATH="/root/workspace/mmdeploy/build/lib:/opt/intel/openvino/deployment_tools/ngraph/lib:/opt/intel/openvino/deployment_tools/inference_engine/lib/intel64:${LD_LIBRARY_PATH}"
ENV LD_LIBRARY_PATH="/root/workspace/mmdeploy/build/lib:${LD_LIBRARY_PATH}"
RUN cd mmdeploy && rm -rf build/CM* && mkdir -p build && cd build && cmake .. \
-DMMDEPLOY_BUILD_SDK=ON \
-DMMDEPLOY_BUILD_EXAMPLES=ON \
-DCMAKE_CXX_COMPILER=g++-7 \
-DCMAKE_CXX_COMPILER=g++-9 \
-DONNXRUNTIME_DIR=${ONNXRUNTIME_DIR} \
-Dncnn_DIR=/root/workspace/ncnn/build/install/lib/cmake/ncnn \
-DInferenceEngine_DIR=/opt/intel/openvino/deployment_tools/inference_engine/share \
-DInferenceEngine_DIR=/opt/intel/openvino/runtime/cmake \
-DMMDEPLOY_TARGET_DEVICES=cpu \
-DMMDEPLOY_BUILD_SDK_PYTHON_API=ON \
-DMMDEPLOY_TARGET_BACKENDS="ort;ncnn;openvino" \
Expand Down