[NVIDIA] Failed inference on OMZ models

[dkurt]

Hi! I did some analysis of OMZ models and found the following issues using NVIDIA plugin. Hope it might be helpful.

Model name	Error
asl-recognition-0004 common-sign-language-0002	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = Sqrt_4372, description = Sqrt; Is not found in OperationRegistry
bert-large-uncased-whole-word-masking-squad-0001 bert-large-uncased-whole-word-masking-squad-emb-0001 bert-small-uncased-whole-word-masking-squad-0001 bert-small-uncased-whole-word-masking-squad-0002	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/cuda_plugin.cpp:74(LoadExeNetworkImpl): Input format 72 is not supported yet. Supported formats are: FP32, FP16, I32, I16, I8, U8 and BOOL.
face-detection-0205 face-detection-0206 horizontal-text-detection-0001 instance-segmentation-person-0007 instance-segmentation-security-0002 instance-segmentation-security-0091 instance-segmentation-security-0228 instance-segmentation-security-1039 instance-segmentation-security-1040 machine-translation-nar-de-en-0002 machine-translation-nar-en-de-0002 machine-translation-nar-en-ru-0002 machine-translation-nar-ru-en-0002 person-detection-0106 person-detection-0203 person-detection-0301 person-detection-0302 person-detection-0303 person-vehicle-bike-detection-2003 person-vehicle-bike-detection-2004 semantic-segmentation-adas-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/cuda_plugin.cpp:60(LoadExeNetworkImpl): Output format 72 is not supported yet. Supported formats are: FP32, FP16, I32, I16, I8, U8 and BOOL.
facial-landmarks-98-detection-0001 human-pose-estimation-0005 human-pose-estimation-0006 human-pose-estimation-0007 icnet-camvid-ava-0001 icnet-camvid-ava-sparse-30-0001 icnet-camvid-ava-sparse-60-0001 road-segmentation-adas-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/interpolate.cpp:62(interpolateFactory): Interpolate node is not supported: not implemented.
faster-rcnn-resnet101-coco-sparse-60-0001 person-detection-retail-0002	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = Proposal_4852, description = Proposal; Is not found in OperationRegistry
human-pose-estimation-0001 person-vehicle-bike-detection-crossroad-0078 time-series-forecasting-electricity-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = Elu_2781, description = Elu; Is not found in OperationRegistry
license-plate-recognition-barrier-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = Tile_2031, description = Tile; Is not found in OperationRegistry
person-detection-action-recognition-0005 person-detection-action-recognition-teacher-0002 person-detection-raisinghand-recognition-0001 person-detection-action-recognition-0006	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = NormalizeL2_9411, description = NormalizeL2; Is not found in OperationRegistry
person-detection-asl-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/cuda_executable_network.cpp:59(ExecutableNetwork): Standard exception from compilation library: get_shape was called on a descriptor::Tensor with dynamic shape
person-vehicle-bike-detection-crossroad-yolov3-1020 yolo-v2-tiny-ava-0001 yolo-v2-tiny-ava-sparse-30-0001 yolo-v2-tiny-ava-sparse-60-0001 yolo-v2-tiny-vehicle-detection-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = RegionYolo_4660, description = RegionYolo; Is not found in OperationRegistry
smartlab-sequence-modelling-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = HardSigmoid_2960, description = HardSigmoid; Is not found in OperationRegistry
unet-camvid-onnx-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = Exp_2112, description = Exp; Is not found in OperationRegistry
yolo-v2-ava-0001 yolo-v2-ava-sparse-35-0001 yolo-v2-ava-sparse-70-0001	[ ERROR ] /content/openvino_contrib/modules/nvidia_plugin/src/ops/subgraph.cpp:59(initExecuteSequence): Node: name = ExtractImagePatches_2200, description = ExtractImagePatches; Is not found in OperationRegistry

[nkogteva]

Please check NVIDIA plugin limitations. List of supported operations https://github.com/openvinotoolkit/openvino_contrib/blob/master/modules/nvidia_plugin/docs/cuda_opset.md