EfficientDet in Deepstream Causes a Seg Fault

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU): Jetson
• DeepStream Version: 6.0
• JetPack Version (valid for Jetson only): 4.6
• TensorRT Version: 8.0.1
• NVIDIA GPU Driver Version (valid for GPU only): N/A
• Issue Type( questions, new requirements, bugs): questions/bugs

Hi,

I trained a grayscale EfficientDet model using TAO. I took the .etlt file over to a Jetson NX and converted it to an engine file using the tao-converter. When I run the deepstream-app, I get the error below. Essentially Deepstream gives a seg fault due to a dimension mismatch. I was first wanting to make sure that EfficientDet works with 1 channel grayscale (according to the TAO docs, this should be ok). I was also wondering what can be fixed in order to make sure the dimensions match.

Thank you,

Megan

clear; ./run.sh
Unknown or legacy key specified 'parse-bbox-func' for group [property]

Using winsys: x11 
gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream/lib/libnvds_nvmultiobjecttracker.so
gstnvtracker: Batch processing is ON
gstnvtracker: Past frame output is ON
[NvMultiObjectTracker] Initialized
0:00:07.008966816 12986     0x13e24630 INFO                 nvinfer gstnvinfer.cpp:638:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:1905> [UID = 1]: deserialized trt engine from :/opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/model/grayscale_efficientdet/model.engine
INFO: [Implicit Engine Info]: layers num: 5
0   INPUT  kFLOAT image_arrays:0  1376x1024x3     
1   OUTPUT kINT32 num_detections  1               
2   OUTPUT kFLOAT detection_boxes 100x4           
3   OUTPUT kFLOAT detection_scores 100             
4   OUTPUT kINT32 detection_classes 100             

WARNING: Backend context bufferIdx(0) request dims:1x1376x1024x1 is out of range, [min: 1x1376x1024x3, max: 1x1376x1024x3]
0:00:07.009377536 12986     0x13e24630 WARN                 nvinfer gstnvinfer.cpp:635:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Warning from NvDsInferContextImpl::checkBackendParams() <nvdsinfer_context_impl.cpp:1862> [UID = 1]: backend can not support dims:1376x1024x1
0:00:07.009471328 12986     0x13e24630 WARN                 nvinfer gstnvinfer.cpp:635:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Warning from NvDsInferContextImpl::generateBackendContext() <nvdsinfer_context_impl.cpp:2017> [UID = 1]: deserialized backend context :/opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/model/grayscale_efficientdet/model.engine failed to match config params, trying rebuild
0:00:07.155242624 12986     0x13e24630 INFO                 nvinfer gstnvinfer.cpp:638:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1919> [UID = 1]: Trying to create engine from model files
WARNING: [TRT]: onnx2trt_utils.cpp:364: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
ERROR: [TRT]: 4: [network.cpp::validate::2724] Error Code 4: Internal Error (image_arrays:0: for dimension number 3 in profile 0 does not match network definition (got min=1, opt=1, max=1), expected min=opt=max=3).)
ERROR: Build engine failed from config file
./run.sh: line 7: 12986 Segmentation fault      (core dumped) deepstream-app -c /opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/configs/main_config.txt

infer_config.txt (886 Bytes)

Moved to TAO Forum.

1 Like

How did you run tao-converter? Can you share the log and the command?

@Morganh

I ran the tao-converter with the following command:

./tao-converter -k nvidia-tlt -c /opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/model/grayscale_efficientdet/model.step-325832.cal image_arrays:0,1x1376x1024x1,8x1376x1024x1,16x1376x1024x1 -e /opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/model/grayscale_efficientdet/model.engine -t int8 -b 8 /opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/model/grayscale_efficientdet/model.etlt

In your config file,

model-engine-file=model.engine

And also in the log

[NvMultiObjectTracker] Initialized
0:00:07.008966816 12986 0x13e24630 INFO nvinfer gstnvinfer.cpp:638:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::deserializeEngineAndBackend() <nvdsinfer_context_impl.cpp:1905> [UID = 1]: deserialized trt engine from :/opt/nvidia/deepstream/deepstream-6.0/samples/models/model_test/model/grayscale_efficientdet/model.engine

Can you inspect this model via below step?

$ python -m pip install colored
$ python -m pip install polygraphy --index-url https://pypi.ngc.nvidia.com
$ polygraphy inspect model model.engine

Hi @Morganh,

Sorry for the delay in getting back to you. Below is the error I receive when running polygraphy:

~Desktop$ polygraphy inspect model model.engine 

[I] Loading bytes from /home/user/Desktop/model.engine
[07/07/2022-13:36:26] [TRT] [E] 1: [stdArchiveReader.cpp::StdArchiveReader::35] Error Code 1: Serialization (Serialization assertion safeVersionRead == safeSerializationVersion failed.Version tag does not match. Note: Current Version: 0, Serialized Engine Version: 43)
[07/07/2022-13:36:26] [TRT] [E] 4: [runtime.cpp::deserializeCudaEngine::50] Error Code 4: Internal Error (Engine deserialization failed.)
[!] Could not deserialize engine. See log for details.

It is due to mismatching version of Tensorrt. Could you please run below steps again in the exact environment where you generate the model.engine?
$ python -m pip install colored
$ python -m pip install polygraphy --index-url https://pypi.ngc.nvidia.com
$ polygraphy inspect model model.engine

Hi @Morganh ,

Below is what was output for the EfficientDet model.

polygraphy inspect model ~/Desktop/model.engine  

[I] Loading bytes from /home/nvidia/Desktop/model.engine
[I] ==== TensorRT Engine ====
    Name: Unnamed Network 0 | Explicit Batch Engine
    
    ---- 1 Engine Input(s) ----
    {image_arrays:0 [dtype=float32, shape=(1, 1376, 1024, 3)]}
    
    ---- 4 Engine Output(s) ----
    {num_detections [dtype=int32, shape=(1, 1)],
     detection_boxes [dtype=float32, shape=(1, 100, 4)],
     detection_scores [dtype=float32, shape=(1, 100)],
     detection_classes [dtype=int32, shape=(1, 100)]}
    
    ---- Memory ----
    Device Memory: 75849728 bytes
    
    ---- 1 Profile(s) (5 Binding(s) Each) ----
    - Profile: 0
        Binding Index: 0 (Input)  [Name: image_arrays:0]    | Shapes: min=(1, 1376, 1024, 3), opt=(1, 1376, 1024, 3), max=(1, 1376, 1024, 3)
        Binding Index: 1 (Output) [Name: num_detections]    | Shape: (1, 1)
        Binding Index: 2 (Output) [Name: detection_boxes]   | Shape: (1, 100, 4)
        Binding Index: 3 (Output) [Name: detection_scores]  | Shape: (1, 100)
        Binding Index: 4 (Output) [Name: detection_classes] | Shape: (1, 100)
    
    ---- 675 Layer(s) ----


So, it does not match the shape when you run tao-converter.
Could you run tao-converter again. Please add "-p ".
i.e.,
-p image_arrays:0,1x1376x1024x1,8x1376x1024x1,16x1376x1024x1

Alright, I added the -p and ran the tao-converter, but I’m still seeing the same thing (e.g. the shape does not match). I’m wondering if there is something going on that is forcing the model to get converted with 3 channels and as a fp32? It seems like no matter what I specify in the tao-converter, I always get a 3 channel, fp32 engine file.

Below is all of the output.

Tao-Converter:

$ ./tao-converter -k nvidia_tlt -p image_arrays:0,1x1376x1024x1,8x1376x1024x1,16x1376x1024x1 -c /opt/nvidia/deepstream/deepstream-6.0/samples/models/test_model/grayscale_efficientdet/model.step-325832.cal -e /opt/nvidia/deepstream/deepstream-6.0/samples/models/test_model/grayscale_efficientdet/model.engine -b 2 -t int8 /opt/nvidia/deepstream/deepstream-6.0/samples/models/test_model/grayscale_efficientdet/model.etlt
[INFO] [MemUsageChange] Init CUDA: CPU +353, GPU +0, now: CPU 371, GPU 5777 (MiB)
[INFO] ----------------------------------------------------------------
[INFO] Input filename:   /tmp/fileXP3R1y
[INFO] ONNX IR version:  0.0.7
[INFO] Opset version:    13
[INFO] Producer name:    
[INFO] Producer version: 
[INFO] Domain:           
[INFO] Model version:    0
[INFO] Doc string:       
[INFO] ----------------------------------------------------------------
[WARNING] onnx2trt_utils.cpp:364: Your ONNX model has been generated with INT64 weights, while TensorRT does not natively support INT64. Attempting to cast down to INT32.
[INFO] No importer registered for op: EfficientNMS_TRT. Attempting to import as plugin.
[INFO] Searching for plugin: EfficientNMS_TRT, plugin_version: 1, plugin_namespace: 
[INFO] Successfully created plugin: EfficientNMS_TRT
[INFO] Detected input dimensions from the model: (1, 1376, 1024, 3)
[INFO] Model has no dynamic shape.
[INFO] [MemUsageSnapshot] Builder begin: CPU 395 MiB, GPU 5840 MiB
[INFO] Reading Calibration Cache for calibrator: EntropyCalibration2
[INFO] Generated calibration scales using calibration cache. Make sure that calibration cache has latest scales.
[INFO] To regenerate calibration cache, please delete the existing one. TensorRT will generate a new calibration cache.
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 97) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 208) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 211) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 309) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 312) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 316) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 319) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 326) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 333) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 339) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 351) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 356) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 362) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 374) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 379) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 385) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 397) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 402) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 408) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 420) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 423) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 430) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 442) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 445) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 448) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 456) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 468) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 471) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 474) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 482) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 494) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 497) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 500) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 508) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 520) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 525) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 531) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 543) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 548) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 554) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 566) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 571) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 577) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 589) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 594) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 600) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 612) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 615) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 622) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 634) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 637) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 640) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 648) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 660) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 663) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 666) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 674) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 686) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 689) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 692) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 700) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 712) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 717) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 723) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 735) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 740) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 746) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 758) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 763) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 769) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 781) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 786) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 792) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 807) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 821) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 854) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 876) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 909) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 931) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 964) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[WARNING] Missing scale and zero-point for tensor (Unnamed Layer* 986) [Constant]_output, expect fall back to non-int8 implementation for any layer consuming or producing given tensor
[INFO] ---------- Layers Running on DLA ----------
[INFO] ---------- Layers Running on GPU ----------
[INFO] [GpuLayer] preprocessor/transpose
[INFO] [GpuLayer] preprocessor/scale_value:0 + preprocessor/scale + preprocessor/mean_value:0 + preprocessor/mean
[INFO] [GpuLayer] stem_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(stem_activation/Sigmoid), stem_activation/mul)
[INFO] [GpuLayer] block1a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block1a_activation/Sigmoid), block1a_activation/mul)
[INFO] [GpuLayer] block1a_se_squeeze/Mean
[INFO] [GpuLayer] block1a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block1a_se_reduce/Sigmoid), block1a_se_reduce/mul)
[INFO] [GpuLayer] block1a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block1a_se_expand/Sigmoid), block1a_se_excite/mul)
[INFO] [GpuLayer] block1a_project_conv/Conv2D
[INFO] [GpuLayer] block2a_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block2a_expand_activation/Sigmoid), block2a_expand_activation/mul)
[INFO] [GpuLayer] block2a_dwconv_pad/Pad + block2a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block2a_activation/Sigmoid), block2a_activation/mul)
[INFO] [GpuLayer] block2a_se_squeeze/Mean
[INFO] [GpuLayer] block2a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block2a_se_reduce/Sigmoid), block2a_se_reduce/mul)
[INFO] [GpuLayer] block2a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block2a_se_expand/Sigmoid), block2a_se_excite/mul)
[INFO] [GpuLayer] block2a_project_conv/Conv2D
[INFO] [GpuLayer] block2b_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block2b_expand_activation/Sigmoid), block2b_expand_activation/mul)
[INFO] [GpuLayer] block2b_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block2b_activation/Sigmoid), block2b_activation/mul)
[INFO] [GpuLayer] block2b_se_squeeze/Mean
[INFO] [GpuLayer] block2b_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block2b_se_reduce/Sigmoid), block2b_se_reduce/mul)
[INFO] [GpuLayer] block2b_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block2b_se_expand/Sigmoid), block2b_se_excite/mul)
[INFO] [GpuLayer] block2b_project_conv/Conv2D + block2b_add/add
[INFO] [GpuLayer] block3a_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block3a_expand_activation/Sigmoid), block3a_expand_activation/mul)
[INFO] [GpuLayer] block3a_dwconv_pad/Pad + block3a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block3a_activation/Sigmoid), block3a_activation/mul)
[INFO] [GpuLayer] block3a_se_squeeze/Mean
[INFO] [GpuLayer] block3a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block3a_se_reduce/Sigmoid), block3a_se_reduce/mul)
[INFO] [GpuLayer] block3a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block3a_se_expand/Sigmoid), block3a_se_excite/mul)
[INFO] [GpuLayer] block3a_project_conv/Conv2D
[INFO] [GpuLayer] block3b_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block3b_expand_activation/Sigmoid), block3b_expand_activation/mul)
[INFO] [GpuLayer] block3b_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block3b_activation/Sigmoid), block3b_activation/mul)
[INFO] [GpuLayer] block3b_se_squeeze/Mean
[INFO] [GpuLayer] block3b_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block3b_se_reduce/Sigmoid), block3b_se_reduce/mul)
[INFO] [GpuLayer] block3b_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block3b_se_expand/Sigmoid), block3b_se_excite/mul)
[INFO] [GpuLayer] block3b_project_conv/Conv2D + block3b_add/add
[INFO] [GpuLayer] bifpn0_0_8_0/BiasAdd || block4a_expand_conv/Conv2D
[INFO] [GpuLayer] bifpn0_0_8_0_bn/FusedBatchNormV3__1956
[INFO] [GpuLayer] PWN(PWN(block4a_expand_activation/Sigmoid), block4a_expand_activation/mul)
[INFO] [GpuLayer] weighted_fusion_3_0/strided_slice__990:0 + (Unnamed Layer* 98) [Shuffle] + weighted_fusion_3_0/mul
[INFO] [GpuLayer] weighted_fusion_3_0/Sum/input_Unsqueeze__1958
[INFO] [GpuLayer] block4a_dwconv_pad/Pad + block4a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block4a_activation/Sigmoid), block4a_activation/mul)
[INFO] [GpuLayer] block4a_se_squeeze/Mean
[INFO] [GpuLayer] block4a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block4a_se_reduce/Sigmoid), block4a_se_reduce/mul)
[INFO] [GpuLayer] block4a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block4a_se_expand/Sigmoid), block4a_se_excite/mul)
[INFO] [GpuLayer] block4a_project_conv/Conv2D
[INFO] [GpuLayer] block4b_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block4b_expand_activation/Sigmoid), block4b_expand_activation/mul)
[INFO] [GpuLayer] block4b_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block4b_activation/Sigmoid), block4b_activation/mul)
[INFO] [GpuLayer] block4b_se_squeeze/Mean
[INFO] [GpuLayer] block4b_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block4b_se_reduce/Sigmoid), block4b_se_reduce/mul)
[INFO] [GpuLayer] block4b_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block4b_se_expand/Sigmoid), block4b_se_excite/mul)
[INFO] [GpuLayer] block4b_project_conv/Conv2D + block4b_add/add
[INFO] [GpuLayer] block4c_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block4c_expand_activation/Sigmoid), block4c_expand_activation/mul)
[INFO] [GpuLayer] block4c_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block4c_activation/Sigmoid), block4c_activation/mul)
[INFO] [GpuLayer] block4c_se_squeeze/Mean
[INFO] [GpuLayer] block4c_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block4c_se_reduce/Sigmoid), block4c_se_reduce/mul)
[INFO] [GpuLayer] block4c_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block4c_se_expand/Sigmoid), block4c_se_excite/mul)
[INFO] [GpuLayer] block4c_project_conv/Conv2D + block4c_add/add
[INFO] [GpuLayer] block5a_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block5a_expand_activation/Sigmoid), block5a_expand_activation/mul)
[INFO] [GpuLayer] block5a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block5a_activation/Sigmoid), block5a_activation/mul)
[INFO] [GpuLayer] block5a_se_squeeze/Mean
[INFO] [GpuLayer] block5a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block5a_se_reduce/Sigmoid), block5a_se_reduce/mul)
[INFO] [GpuLayer] block5a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block5a_se_expand/Sigmoid), block5a_se_excite/mul)
[INFO] [GpuLayer] block5a_project_conv/Conv2D
[INFO] [GpuLayer] block5b_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block5b_expand_activation/Sigmoid), block5b_expand_activation/mul)
[INFO] [GpuLayer] block5b_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block5b_activation/Sigmoid), block5b_activation/mul)
[INFO] [GpuLayer] block5b_se_squeeze/Mean
[INFO] [GpuLayer] block5b_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block5b_se_reduce/Sigmoid), block5b_se_reduce/mul)
[INFO] [GpuLayer] block5b_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block5b_se_expand/Sigmoid), block5b_se_excite/mul)
[INFO] [GpuLayer] block5b_project_conv/Conv2D + block5b_add/add
[INFO] [GpuLayer] block5c_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block5c_expand_activation/Sigmoid), block5c_expand_activation/mul)
[INFO] [GpuLayer] block5c_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block5c_activation/Sigmoid), block5c_activation/mul)
[INFO] [GpuLayer] block5c_se_squeeze/Mean
[INFO] [GpuLayer] block5c_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block5c_se_reduce/Sigmoid), block5c_se_reduce/mul)
[INFO] [GpuLayer] block5c_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block5c_se_expand/Sigmoid), block5c_se_excite/mul)
[INFO] [GpuLayer] block5c_project_conv/Conv2D + block5c_add/add
[INFO] [GpuLayer] bifpn0_1_9_0/BiasAdd || bifpn0_1_7_0/BiasAdd || block6a_expand_conv/Conv2D
[INFO] [GpuLayer] bifpn0_1_9_0_bn/FusedBatchNormV3__1921
[INFO] [GpuLayer] bifpn0_1_7_0_bn/FusedBatchNormV3__1925
[INFO] [GpuLayer] PWN(PWN(block6a_expand_activation/Sigmoid), block6a_expand_activation/mul)
[INFO] [GpuLayer] weighted_fusion_4_0/strided_slice__957:0 + (Unnamed Layer* 209) [Shuffle] + weighted_fusion_4_0/mul
[INFO] [GpuLayer] weighted_fusion_2_0/strided_slice__1023:0 + (Unnamed Layer* 212) [Shuffle] + weighted_fusion_2_0/mul
[INFO] [GpuLayer] weighted_fusion_4_0/Sum/input_Unsqueeze__1977
[INFO] [GpuLayer] weighted_fusion_2_0/Sum/input_Unsqueeze__1927
[INFO] [GpuLayer] block6a_dwconv_pad/Pad + block6a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block6a_activation/Sigmoid), block6a_activation/mul)
[INFO] [GpuLayer] block6a_se_squeeze/Mean
[INFO] [GpuLayer] block6a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6a_se_reduce/Sigmoid), block6a_se_reduce/mul)
[INFO] [GpuLayer] block6a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6a_se_expand/Sigmoid), block6a_se_excite/mul)
[INFO] [GpuLayer] block6a_project_conv/Conv2D
[INFO] [GpuLayer] block6b_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block6b_expand_activation/Sigmoid), block6b_expand_activation/mul)
[INFO] [GpuLayer] block6b_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block6b_activation/Sigmoid), block6b_activation/mul)
[INFO] [GpuLayer] block6b_se_squeeze/Mean
[INFO] [GpuLayer] block6b_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6b_se_reduce/Sigmoid), block6b_se_reduce/mul)
[INFO] [GpuLayer] block6b_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6b_se_expand/Sigmoid), block6b_se_excite/mul)
[INFO] [GpuLayer] block6b_project_conv/Conv2D + block6b_add/add
[INFO] [GpuLayer] block6c_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block6c_expand_activation/Sigmoid), block6c_expand_activation/mul)
[INFO] [GpuLayer] block6c_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block6c_activation/Sigmoid), block6c_activation/mul)
[INFO] [GpuLayer] block6c_se_squeeze/Mean
[INFO] [GpuLayer] block6c_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6c_se_reduce/Sigmoid), block6c_se_reduce/mul)
[INFO] [GpuLayer] block6c_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6c_se_expand/Sigmoid), block6c_se_excite/mul)
[INFO] [GpuLayer] block6c_project_conv/Conv2D + block6c_add/add
[INFO] [GpuLayer] block6d_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block6d_expand_activation/Sigmoid), block6d_expand_activation/mul)
[INFO] [GpuLayer] block6d_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block6d_activation/Sigmoid), block6d_activation/mul)
[INFO] [GpuLayer] block6d_se_squeeze/Mean
[INFO] [GpuLayer] block6d_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6d_se_reduce/Sigmoid), block6d_se_reduce/mul)
[INFO] [GpuLayer] block6d_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block6d_se_expand/Sigmoid), block6d_se_excite/mul)
[INFO] [GpuLayer] block6d_project_conv/Conv2D + block6d_add/add
[INFO] [GpuLayer] block7a_expand_conv/Conv2D
[INFO] [GpuLayer] PWN(PWN(block7a_expand_activation/Sigmoid), block7a_expand_activation/mul)
[INFO] [GpuLayer] block7a_dwconv/depthwise
[INFO] [GpuLayer] PWN(PWN(block7a_activation/Sigmoid), block7a_activation/mul)
[INFO] [GpuLayer] block7a_se_squeeze/Mean
[INFO] [GpuLayer] block7a_se_reduce/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block7a_se_reduce/Sigmoid), block7a_se_reduce/mul)
[INFO] [GpuLayer] block7a_se_expand/BiasAdd
[INFO] [GpuLayer] PWN(PWN(block7a_se_expand/Sigmoid), block7a_se_excite/mul)
[INFO] [GpuLayer] block7a_project_conv/Conv2D
[INFO] [GpuLayer] p6/BiasAdd || bifpn0_2_6_0/BiasAdd || bifpn0_2_10_0/BiasAdd
[INFO] [GpuLayer] max_pooling2d/MaxPool
[INFO] [GpuLayer] bifpn0_2_6_0_bn/FusedBatchNormV3__1886
[INFO] [GpuLayer] bifpn0_2_10_0_bn/FusedBatchNormV3__1917
[INFO] [GpuLayer] max_pooling2d_1/MaxPool
[INFO] [GpuLayer] Transpose__3595
[INFO] [GpuLayer] weighted_fusion_1_0/strided_slice__1056:0 + (Unnamed Layer* 310) [Shuffle] + weighted_fusion_1_0/mul
[INFO] [GpuLayer] weighted_fusion_5_0/strided_slice__909:0 + (Unnamed Layer* 313) [Shuffle] + weighted_fusion_5_0/mul
[INFO] [GpuLayer] Transpose__3046
[INFO] [GpuLayer] Resize__1855
[INFO] [GpuLayer] weighted_fusion_6_0/strided_slice__861:0 + (Unnamed Layer* 317) [Shuffle] + weighted_fusion_6_0/mul
[INFO] [GpuLayer] weighted_fusion_0_0/strided_slice__1089:0 + (Unnamed Layer* 320) [Shuffle] + weighted_fusion_0_0/mul
[INFO] [GpuLayer] weighted_fusion_1_0/Sum/input_Unsqueeze__1888
[INFO] [GpuLayer] weighted_fusion_5_0/Sum/input_Unsqueeze__1998
[INFO] [GpuLayer] weighted_fusion_7_0/strided_slice__813:0 + (Unnamed Layer* 327) [Shuffle] + weighted_fusion_7_0/mul
[INFO] [GpuLayer] image_resize_layer/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_6_0/Sum/input_Unsqueeze__2019
[INFO] [GpuLayer] weighted_fusion_0_0/Sum/input_Unsqueeze__1857
[INFO] [GpuLayer] weighted_fusion_7_0/Sum/input_Unsqueeze__2040
[INFO] [GpuLayer] weighted_fusion_0_0/strided_slice_1__1084:0 + (Unnamed Layer* 334) [Shuffle] + weighted_fusion_0_0/mul_1
[INFO] [GpuLayer] weighted_fusion_0_0/Sum/input_Unsqueeze__1859
[INFO] [GpuLayer] weighted_fusion_0_0/Sum/input_Unsqueeze__1857:0 copy
[INFO] [GpuLayer] weighted_fusion_0_0/Sum/input_Unsqueeze__1859:0 copy
[INFO] [GpuLayer] weighted_fusion_0_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_0_0/add:0 + (Unnamed Layer* 340) [Shuffle], weighted_fusion_0_0/truediv), PWN(PWN(activation/Sigmoid), activation/mul))
[INFO] [GpuLayer] after_combine_dw_conv_0_0/separable_conv2d/depthwise__1864
[INFO] [GpuLayer] after_combine_dw_conv_0_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_0_0/BiasAdd
[INFO] [GpuLayer] Transpose__3047
[INFO] [GpuLayer] Resize__1882
[INFO] [GpuLayer] weighted_fusion_6_0/strided_slice_1__856:0 + (Unnamed Layer* 352) [Shuffle] + weighted_fusion_6_0/mul_1
[INFO] [GpuLayer] image_resize_layer_1/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_6_0/Sum/input_Unsqueeze__2021
[INFO] [GpuLayer] weighted_fusion_1_0/strided_slice_1__1051:0 + (Unnamed Layer* 357) [Shuffle] + weighted_fusion_1_0/mul_1
[INFO] [GpuLayer] weighted_fusion_1_0/Sum/input_Unsqueeze__1890
[INFO] [GpuLayer] weighted_fusion_1_0/Sum/input_Unsqueeze__1888:0 copy
[INFO] [GpuLayer] weighted_fusion_1_0/Sum/input_Unsqueeze__1890:0 copy
[INFO] [GpuLayer] weighted_fusion_1_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_1_0/add:0 + (Unnamed Layer* 363) [Shuffle], weighted_fusion_1_0/truediv), PWN(PWN(activation_1/Sigmoid), activation_1/mul))
[INFO] [GpuLayer] after_combine_dw_conv_1_0/separable_conv2d/depthwise__1895
[INFO] [GpuLayer] after_combine_dw_conv_1_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_1_0/BiasAdd
[INFO] [GpuLayer] Transpose__3050
[INFO] [GpuLayer] Resize__1913
[INFO] [GpuLayer] weighted_fusion_5_0/strided_slice_1__904:0 + (Unnamed Layer* 375) [Shuffle] + weighted_fusion_5_0/mul_1
[INFO] [GpuLayer] image_resize_layer_2/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_5_0/Sum/input_Unsqueeze__2000
[INFO] [GpuLayer] weighted_fusion_2_0/strided_slice_1__1018:0 + (Unnamed Layer* 380) [Shuffle] + weighted_fusion_2_0/mul_1
[INFO] [GpuLayer] weighted_fusion_2_0/Sum/input_Unsqueeze__1929
[INFO] [GpuLayer] weighted_fusion_2_0/Sum/input_Unsqueeze__1927:0 copy
[INFO] [GpuLayer] weighted_fusion_2_0/Sum/input_Unsqueeze__1929:0 copy
[INFO] [GpuLayer] weighted_fusion_2_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_2_0/add:0 + (Unnamed Layer* 386) [Shuffle], weighted_fusion_2_0/truediv), PWN(PWN(activation_2/Sigmoid), activation_2/mul))
[INFO] [GpuLayer] after_combine_dw_conv_2_0/separable_conv2d/depthwise__1934
[INFO] [GpuLayer] after_combine_dw_conv_2_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_2_0/BiasAdd
[INFO] [GpuLayer] Transpose__3052
[INFO] [GpuLayer] Resize__1952
[INFO] [GpuLayer] weighted_fusion_4_0/strided_slice_1__952:0 + (Unnamed Layer* 398) [Shuffle] + weighted_fusion_4_0/mul_1
[INFO] [GpuLayer] image_resize_layer_3/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_4_0/Sum/input_Unsqueeze__1979
[INFO] [GpuLayer] weighted_fusion_3_0/strided_slice_1__985:0 + (Unnamed Layer* 403) [Shuffle] + weighted_fusion_3_0/mul_1
[INFO] [GpuLayer] weighted_fusion_3_0/Sum/input_Unsqueeze__1960
[INFO] [GpuLayer] weighted_fusion_3_0/Sum/input_Unsqueeze__1958:0 copy
[INFO] [GpuLayer] weighted_fusion_3_0/Sum/input_Unsqueeze__1960:0 copy
[INFO] [GpuLayer] weighted_fusion_3_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_3_0/add:0 + (Unnamed Layer* 409) [Shuffle], weighted_fusion_3_0/truediv), PWN(PWN(activation_3/Sigmoid), activation_3/mul))
[INFO] [GpuLayer] after_combine_dw_conv_3_0/separable_conv2d/depthwise__1965
[INFO] [GpuLayer] after_combine_dw_conv_3_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_3_0/BiasAdd
[INFO] [GpuLayer] max_pooling2d_2/MaxPool
[INFO] [GpuLayer] Transpose__3053
[INFO] [GpuLayer] max_pooling2d_2/MaxPool__1975
[INFO] [GpuLayer] weighted_fusion_3_1/strided_slice__979:0 + (Unnamed Layer* 421) [Shuffle] + weighted_fusion_3_1/mul
[INFO] [GpuLayer] weighted_fusion_4_0/strided_slice_2__947:0 + (Unnamed Layer* 424) [Shuffle] + weighted_fusion_4_0/mul_2
[INFO] [GpuLayer] weighted_fusion_3_1/Sum/input_Unsqueeze__2148
[INFO] [GpuLayer] weighted_fusion_4_0/Sum/input_Unsqueeze__1981
[INFO] [GpuLayer] weighted_fusion_4_0/Sum/input_Unsqueeze__1977:0 copy
[INFO] [GpuLayer] weighted_fusion_4_0/Sum/input_Unsqueeze__1979:0 copy
[INFO] [GpuLayer] weighted_fusion_4_0/Sum/input_Unsqueeze__1981:0 copy
[INFO] [GpuLayer] weighted_fusion_4_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_4_0/add:0 + (Unnamed Layer* 431) [Shuffle], weighted_fusion_4_0/truediv), PWN(PWN(activation_4/Sigmoid), activation_4/mul))
[INFO] [GpuLayer] after_combine_dw_conv_4_0/separable_conv2d/depthwise__1986
[INFO] [GpuLayer] after_combine_dw_conv_4_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_4_0/BiasAdd
[INFO] [GpuLayer] max_pooling2d_3/MaxPool
[INFO] [GpuLayer] Transpose__3603
[INFO] [GpuLayer] max_pooling2d_3/MaxPool__1996
[INFO] [GpuLayer] weighted_fusion_4_1/strided_slice__941:0 + (Unnamed Layer* 443) [Shuffle] + weighted_fusion_4_1/mul
[INFO] [GpuLayer] weighted_fusion_2_1/strided_slice__1012:0 + (Unnamed Layer* 446) [Shuffle] + weighted_fusion_2_1/mul
[INFO] [GpuLayer] weighted_fusion_5_0/strided_slice_2__899:0 + (Unnamed Layer* 449) [Shuffle] + weighted_fusion_5_0/mul_2
[INFO] [GpuLayer] weighted_fusion_4_1/Sum/input_Unsqueeze__2167
[INFO] [GpuLayer] weighted_fusion_2_1/Sum/input_Unsqueeze__2121
[INFO] [GpuLayer] weighted_fusion_5_0/Sum/input_Unsqueeze__2002
[INFO] [GpuLayer] weighted_fusion_5_0/Sum/input_Unsqueeze__1998:0 copy
[INFO] [GpuLayer] weighted_fusion_5_0/Sum/input_Unsqueeze__2000:0 copy
[INFO] [GpuLayer] weighted_fusion_5_0/Sum/input_Unsqueeze__2002:0 copy
[INFO] [GpuLayer] weighted_fusion_5_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_5_0/add:0 + (Unnamed Layer* 457) [Shuffle], weighted_fusion_5_0/truediv), PWN(PWN(activation_5/Sigmoid), activation_5/mul))
[INFO] [GpuLayer] after_combine_dw_conv_5_0/separable_conv2d/depthwise__2007
[INFO] [GpuLayer] after_combine_dw_conv_5_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_5_0/BiasAdd
[INFO] [GpuLayer] max_pooling2d_4/MaxPool
[INFO] [GpuLayer] Transpose__3599
[INFO] [GpuLayer] max_pooling2d_4/MaxPool__2017
[INFO] [GpuLayer] weighted_fusion_5_1/strided_slice__893:0 + (Unnamed Layer* 469) [Shuffle] + weighted_fusion_5_1/mul
[INFO] [GpuLayer] weighted_fusion_1_1/strided_slice__1045:0 + (Unnamed Layer* 472) [Shuffle] + weighted_fusion_1_1/mul
[INFO] [GpuLayer] weighted_fusion_6_0/strided_slice_2__851:0 + (Unnamed Layer* 475) [Shuffle] + weighted_fusion_6_0/mul_2
[INFO] [GpuLayer] weighted_fusion_5_1/Sum/input_Unsqueeze__2188
[INFO] [GpuLayer] weighted_fusion_1_1/Sum/input_Unsqueeze__2094
[INFO] [GpuLayer] weighted_fusion_6_0/Sum/input_Unsqueeze__2023
[INFO] [GpuLayer] weighted_fusion_6_0/Sum/input_Unsqueeze__2019:0 copy
[INFO] [GpuLayer] weighted_fusion_6_0/Sum/input_Unsqueeze__2021:0 copy
[INFO] [GpuLayer] weighted_fusion_6_0/Sum/input_Unsqueeze__2023:0 copy
[INFO] [GpuLayer] weighted_fusion_6_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_6_0/add:0 + (Unnamed Layer* 483) [Shuffle], weighted_fusion_6_0/truediv), PWN(PWN(activation_6/Sigmoid), activation_6/mul))
[INFO] [GpuLayer] after_combine_dw_conv_6_0/separable_conv2d/depthwise__2028
[INFO] [GpuLayer] after_combine_dw_conv_6_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_6_0/BiasAdd
[INFO] [GpuLayer] max_pooling2d_5/MaxPool
[INFO] [GpuLayer] Transpose__3605
[INFO] [GpuLayer] max_pooling2d_5/MaxPool__2038
[INFO] [GpuLayer] weighted_fusion_6_1/strided_slice__845:0 + (Unnamed Layer* 495) [Shuffle] + weighted_fusion_6_1/mul
[INFO] [GpuLayer] weighted_fusion_0_1/strided_slice__1078:0 + (Unnamed Layer* 498) [Shuffle] + weighted_fusion_0_1/mul
[INFO] [GpuLayer] weighted_fusion_7_0/strided_slice_1__808:0 + (Unnamed Layer* 501) [Shuffle] + weighted_fusion_7_0/mul_1
[INFO] [GpuLayer] weighted_fusion_6_1/Sum/input_Unsqueeze__2209
[INFO] [GpuLayer] weighted_fusion_0_1/Sum/input_Unsqueeze__2067
[INFO] [GpuLayer] weighted_fusion_7_0/Sum/input_Unsqueeze__2042
[INFO] [GpuLayer] weighted_fusion_7_0/Sum/input_Unsqueeze__2040:0 copy
[INFO] [GpuLayer] weighted_fusion_7_0/Sum/input_Unsqueeze__2042:0 copy
[INFO] [GpuLayer] weighted_fusion_7_0/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_7_0/add:0 + (Unnamed Layer* 509) [Shuffle], weighted_fusion_7_0/truediv), PWN(PWN(activation_7/Sigmoid), activation_7/mul))
[INFO] [GpuLayer] after_combine_dw_conv_7_0/separable_conv2d/depthwise__2047
[INFO] [GpuLayer] after_combine_dw_conv_7_0/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_7_0/BiasAdd
[INFO] [GpuLayer] Transpose__3065
[INFO] [GpuLayer] Resize__2065
[INFO] [GpuLayer] weighted_fusion_7_1/strided_slice__802:0 + (Unnamed Layer* 521) [Shuffle] + weighted_fusion_7_1/mul
[INFO] [GpuLayer] image_resize_layer_4/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_7_1/Sum/input_Unsqueeze__2230
[INFO] [GpuLayer] weighted_fusion_0_1/strided_slice_1__1073:0 + (Unnamed Layer* 526) [Shuffle] + weighted_fusion_0_1/mul_1
[INFO] [GpuLayer] weighted_fusion_0_1/Sum/input_Unsqueeze__2069
[INFO] [GpuLayer] weighted_fusion_0_1/Sum/input_Unsqueeze__2067:0 copy
[INFO] [GpuLayer] weighted_fusion_0_1/Sum/input_Unsqueeze__2069:0 copy
[INFO] [GpuLayer] weighted_fusion_0_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_0_1/add:0 + (Unnamed Layer* 532) [Shuffle], weighted_fusion_0_1/truediv), PWN(PWN(activation_8/Sigmoid), activation_8/mul))
[INFO] [GpuLayer] after_combine_dw_conv_0_1/separable_conv2d/depthwise__2074
[INFO] [GpuLayer] after_combine_dw_conv_0_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_0_1/BiasAdd
[INFO] [GpuLayer] Transpose__3067
[INFO] [GpuLayer] Resize__2092
[INFO] [GpuLayer] weighted_fusion_6_1/strided_slice_1__840:0 + (Unnamed Layer* 544) [Shuffle] + weighted_fusion_6_1/mul_1
[INFO] [GpuLayer] image_resize_layer_5/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_6_1/Sum/input_Unsqueeze__2211
[INFO] [GpuLayer] weighted_fusion_1_1/strided_slice_1__1040:0 + (Unnamed Layer* 549) [Shuffle] + weighted_fusion_1_1/mul_1
[INFO] [GpuLayer] weighted_fusion_1_1/Sum/input_Unsqueeze__2096
[INFO] [GpuLayer] weighted_fusion_1_1/Sum/input_Unsqueeze__2094:0 copy
[INFO] [GpuLayer] weighted_fusion_1_1/Sum/input_Unsqueeze__2096:0 copy
[INFO] [GpuLayer] weighted_fusion_1_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_1_1/add:0 + (Unnamed Layer* 555) [Shuffle], weighted_fusion_1_1/truediv), PWN(PWN(activation_9/Sigmoid), activation_9/mul))
[INFO] [GpuLayer] after_combine_dw_conv_1_1/separable_conv2d/depthwise__2101
[INFO] [GpuLayer] after_combine_dw_conv_1_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_1_1/BiasAdd
[INFO] [GpuLayer] Transpose__3069
[INFO] [GpuLayer] Resize__2119
[INFO] [GpuLayer] weighted_fusion_5_1/strided_slice_1__888:0 + (Unnamed Layer* 567) [Shuffle] + weighted_fusion_5_1/mul_1
[INFO] [GpuLayer] image_resize_layer_6/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_5_1/Sum/input_Unsqueeze__2190
[INFO] [GpuLayer] weighted_fusion_2_1/strided_slice_1__1007:0 + (Unnamed Layer* 572) [Shuffle] + weighted_fusion_2_1/mul_1
[INFO] [GpuLayer] weighted_fusion_2_1/Sum/input_Unsqueeze__2123
[INFO] [GpuLayer] weighted_fusion_2_1/Sum/input_Unsqueeze__2121:0 copy
[INFO] [GpuLayer] weighted_fusion_2_1/Sum/input_Unsqueeze__2123:0 copy
[INFO] [GpuLayer] weighted_fusion_2_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_2_1/add:0 + (Unnamed Layer* 578) [Shuffle], weighted_fusion_2_1/truediv), PWN(PWN(activation_10/Sigmoid), activation_10/mul))
[INFO] [GpuLayer] after_combine_dw_conv_2_1/separable_conv2d/depthwise__2128
[INFO] [GpuLayer] after_combine_dw_conv_2_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_2_1/BiasAdd
[INFO] [GpuLayer] Transpose__3071
[INFO] [GpuLayer] Resize__2146
[INFO] [GpuLayer] weighted_fusion_4_1/strided_slice_1__936:0 + (Unnamed Layer* 590) [Shuffle] + weighted_fusion_4_1/mul_1
[INFO] [GpuLayer] image_resize_layer_7/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_4_1/Sum/input_Unsqueeze__2169
[INFO] [GpuLayer] weighted_fusion_3_1/strided_slice_1__974:0 + (Unnamed Layer* 595) [Shuffle] + weighted_fusion_3_1/mul_1
[INFO] [GpuLayer] weighted_fusion_3_1/Sum/input_Unsqueeze__2150
[INFO] [GpuLayer] weighted_fusion_3_1/Sum/input_Unsqueeze__2148:0 copy
[INFO] [GpuLayer] weighted_fusion_3_1/Sum/input_Unsqueeze__2150:0 copy
[INFO] [GpuLayer] weighted_fusion_3_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_3_1/add:0 + (Unnamed Layer* 601) [Shuffle], weighted_fusion_3_1/truediv), PWN(PWN(activation_11/Sigmoid), activation_11/mul))
[INFO] [GpuLayer] after_combine_dw_conv_3_1/separable_conv2d/depthwise__2155
[INFO] [GpuLayer] after_combine_dw_conv_3_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_3_1/BiasAdd
[INFO] [GpuLayer] max_pooling2d_6/MaxPool
[INFO] [GpuLayer] Transpose__3072
[INFO] [GpuLayer] max_pooling2d_6/MaxPool__2165
[INFO] [GpuLayer] weighted_fusion_3_2/strided_slice__968:0 + (Unnamed Layer* 613) [Shuffle] + weighted_fusion_3_2/mul
[INFO] [GpuLayer] weighted_fusion_4_1/strided_slice_2__931:0 + (Unnamed Layer* 616) [Shuffle] + weighted_fusion_4_1/mul_2
[INFO] [GpuLayer] weighted_fusion_3_2/Sum/input_Unsqueeze__2338
[INFO] [GpuLayer] weighted_fusion_4_1/Sum/input_Unsqueeze__2171
[INFO] [GpuLayer] weighted_fusion_4_1/Sum/input_Unsqueeze__2167:0 copy
[INFO] [GpuLayer] weighted_fusion_4_1/Sum/input_Unsqueeze__2169:0 copy
[INFO] [GpuLayer] weighted_fusion_4_1/Sum/input_Unsqueeze__2171:0 copy
[INFO] [GpuLayer] weighted_fusion_4_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_4_1/add:0 + (Unnamed Layer* 623) [Shuffle], weighted_fusion_4_1/truediv), PWN(PWN(activation_12/Sigmoid), activation_12/mul))
[INFO] [GpuLayer] after_combine_dw_conv_4_1/separable_conv2d/depthwise__2176
[INFO] [GpuLayer] after_combine_dw_conv_4_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_4_1/BiasAdd
[INFO] [GpuLayer] max_pooling2d_7/MaxPool
[INFO] [GpuLayer] Transpose__3607
[INFO] [GpuLayer] max_pooling2d_7/MaxPool__2186
[INFO] [GpuLayer] weighted_fusion_4_2/strided_slice__925:0 + (Unnamed Layer* 635) [Shuffle] + weighted_fusion_4_2/mul
[INFO] [GpuLayer] weighted_fusion_2_2/strided_slice__1001:0 + (Unnamed Layer* 638) [Shuffle] + weighted_fusion_2_2/mul
[INFO] [GpuLayer] weighted_fusion_5_1/strided_slice_2__883:0 + (Unnamed Layer* 641) [Shuffle] + weighted_fusion_5_1/mul_2
[INFO] [GpuLayer] weighted_fusion_4_2/Sum/input_Unsqueeze__2357
[INFO] [GpuLayer] weighted_fusion_2_2/Sum/input_Unsqueeze__2311
[INFO] [GpuLayer] weighted_fusion_5_1/Sum/input_Unsqueeze__2192
[INFO] [GpuLayer] weighted_fusion_5_1/Sum/input_Unsqueeze__2188:0 copy
[INFO] [GpuLayer] weighted_fusion_5_1/Sum/input_Unsqueeze__2190:0 copy
[INFO] [GpuLayer] weighted_fusion_5_1/Sum/input_Unsqueeze__2192:0 copy
[INFO] [GpuLayer] weighted_fusion_5_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_5_1/add:0 + (Unnamed Layer* 649) [Shuffle], weighted_fusion_5_1/truediv), PWN(PWN(activation_13/Sigmoid), activation_13/mul))
[INFO] [GpuLayer] after_combine_dw_conv_5_1/separable_conv2d/depthwise__2197
[INFO] [GpuLayer] after_combine_dw_conv_5_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_5_1/BiasAdd
[INFO] [GpuLayer] max_pooling2d_8/MaxPool
[INFO] [GpuLayer] Transpose__3597
[INFO] [GpuLayer] max_pooling2d_8/MaxPool__2207
[INFO] [GpuLayer] weighted_fusion_5_2/strided_slice__877:0 + (Unnamed Layer* 661) [Shuffle] + weighted_fusion_5_2/mul
[INFO] [GpuLayer] weighted_fusion_1_2/strided_slice__1034:0 + (Unnamed Layer* 664) [Shuffle] + weighted_fusion_1_2/mul
[INFO] [GpuLayer] weighted_fusion_6_1/strided_slice_2__835:0 + (Unnamed Layer* 667) [Shuffle] + weighted_fusion_6_1/mul_2
[INFO] [GpuLayer] weighted_fusion_5_2/Sum/input_Unsqueeze__2378
[INFO] [GpuLayer] weighted_fusion_1_2/Sum/input_Unsqueeze__2284
[INFO] [GpuLayer] weighted_fusion_6_1/Sum/input_Unsqueeze__2213
[INFO] [GpuLayer] weighted_fusion_6_1/Sum/input_Unsqueeze__2209:0 copy
[INFO] [GpuLayer] weighted_fusion_6_1/Sum/input_Unsqueeze__2211:0 copy
[INFO] [GpuLayer] weighted_fusion_6_1/Sum/input_Unsqueeze__2213:0 copy
[INFO] [GpuLayer] weighted_fusion_6_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_6_1/add:0 + (Unnamed Layer* 675) [Shuffle], weighted_fusion_6_1/truediv), PWN(PWN(activation_14/Sigmoid), activation_14/mul))
[INFO] [GpuLayer] after_combine_dw_conv_6_1/separable_conv2d/depthwise__2218
[INFO] [GpuLayer] after_combine_dw_conv_6_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_6_1/BiasAdd
[INFO] [GpuLayer] max_pooling2d_9/MaxPool
[INFO] [GpuLayer] Transpose__3601
[INFO] [GpuLayer] max_pooling2d_9/MaxPool__2228
[INFO] [GpuLayer] weighted_fusion_6_2/strided_slice__829:0 + (Unnamed Layer* 687) [Shuffle] + weighted_fusion_6_2/mul
[INFO] [GpuLayer] weighted_fusion_0_2/strided_slice__1067:0 + (Unnamed Layer* 690) [Shuffle] + weighted_fusion_0_2/mul
[INFO] [GpuLayer] weighted_fusion_7_1/strided_slice_1__797:0 + (Unnamed Layer* 693) [Shuffle] + weighted_fusion_7_1/mul_1
[INFO] [GpuLayer] weighted_fusion_6_2/Sum/input_Unsqueeze__2399
[INFO] [GpuLayer] weighted_fusion_0_2/Sum/input_Unsqueeze__2257
[INFO] [GpuLayer] weighted_fusion_7_1/Sum/input_Unsqueeze__2232
[INFO] [GpuLayer] weighted_fusion_7_1/Sum/input_Unsqueeze__2230:0 copy
[INFO] [GpuLayer] weighted_fusion_7_1/Sum/input_Unsqueeze__2232:0 copy
[INFO] [GpuLayer] weighted_fusion_7_1/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_7_1/add:0 + (Unnamed Layer* 701) [Shuffle], weighted_fusion_7_1/truediv), PWN(PWN(activation_15/Sigmoid), activation_15/mul))
[INFO] [GpuLayer] after_combine_dw_conv_7_1/separable_conv2d/depthwise__2237
[INFO] [GpuLayer] after_combine_dw_conv_7_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_7_1/BiasAdd
[INFO] [GpuLayer] Transpose__3083
[INFO] [GpuLayer] Resize__2255
[INFO] [GpuLayer] weighted_fusion_7_2/strided_slice__791:0 + (Unnamed Layer* 713) [Shuffle] + weighted_fusion_7_2/mul
[INFO] [GpuLayer] image_resize_layer_8/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_7_2/Sum/input_Unsqueeze__2420
[INFO] [GpuLayer] weighted_fusion_0_2/strided_slice_1__1062:0 + (Unnamed Layer* 718) [Shuffle] + weighted_fusion_0_2/mul_1
[INFO] [GpuLayer] weighted_fusion_0_2/Sum/input_Unsqueeze__2259
[INFO] [GpuLayer] weighted_fusion_0_2/Sum/input_Unsqueeze__2257:0 copy
[INFO] [GpuLayer] weighted_fusion_0_2/Sum/input_Unsqueeze__2259:0 copy
[INFO] [GpuLayer] weighted_fusion_0_2/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_0_2/add:0 + (Unnamed Layer* 724) [Shuffle], weighted_fusion_0_2/truediv), PWN(PWN(activation_16/Sigmoid), activation_16/mul))
[INFO] [GpuLayer] after_combine_dw_conv_0_2/separable_conv2d/depthwise__2264
[INFO] [GpuLayer] after_combine_dw_conv_0_2/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_0_2/BiasAdd
[INFO] [GpuLayer] Transpose__3085
[INFO] [GpuLayer] Resize__2282
[INFO] [GpuLayer] weighted_fusion_6_2/strided_slice_1__824:0 + (Unnamed Layer* 736) [Shuffle] + weighted_fusion_6_2/mul_1
[INFO] [GpuLayer] image_resize_layer_9/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_6_2/Sum/input_Unsqueeze__2401
[INFO] [GpuLayer] weighted_fusion_1_2/strided_slice_1__1029:0 + (Unnamed Layer* 741) [Shuffle] + weighted_fusion_1_2/mul_1
[INFO] [GpuLayer] weighted_fusion_1_2/Sum/input_Unsqueeze__2286
[INFO] [GpuLayer] weighted_fusion_1_2/Sum/input_Unsqueeze__2284:0 copy
[INFO] [GpuLayer] weighted_fusion_1_2/Sum/input_Unsqueeze__2286:0 copy
[INFO] [GpuLayer] weighted_fusion_1_2/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_1_2/add:0 + (Unnamed Layer* 747) [Shuffle], weighted_fusion_1_2/truediv), PWN(PWN(activation_17/Sigmoid), activation_17/mul))
[INFO] [GpuLayer] after_combine_dw_conv_1_2/separable_conv2d/depthwise__2291
[INFO] [GpuLayer] after_combine_dw_conv_1_2/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_1_2/BiasAdd
[INFO] [GpuLayer] Transpose__3087
[INFO] [GpuLayer] Resize__2309
[INFO] [GpuLayer] weighted_fusion_5_2/strided_slice_1__872:0 + (Unnamed Layer* 759) [Shuffle] + weighted_fusion_5_2/mul_1
[INFO] [GpuLayer] image_resize_layer_10/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_5_2/Sum/input_Unsqueeze__2380
[INFO] [GpuLayer] weighted_fusion_2_2/strided_slice_1__996:0 + (Unnamed Layer* 764) [Shuffle] + weighted_fusion_2_2/mul_1
[INFO] [GpuLayer] weighted_fusion_2_2/Sum/input_Unsqueeze__2313
[INFO] [GpuLayer] weighted_fusion_2_2/Sum/input_Unsqueeze__2311:0 copy
[INFO] [GpuLayer] weighted_fusion_2_2/Sum/input_Unsqueeze__2313:0 copy
[INFO] [GpuLayer] weighted_fusion_2_2/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_2_2/add:0 + (Unnamed Layer* 770) [Shuffle], weighted_fusion_2_2/truediv), PWN(PWN(activation_18/Sigmoid), activation_18/mul))
[INFO] [GpuLayer] after_combine_dw_conv_2_2/separable_conv2d/depthwise__2318
[INFO] [GpuLayer] after_combine_dw_conv_2_2/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_2_2/BiasAdd
[INFO] [GpuLayer] Transpose__3090
[INFO] [GpuLayer] Resize__2336
[INFO] [GpuLayer] weighted_fusion_4_2/strided_slice_1__920:0 + (Unnamed Layer* 782) [Shuffle] + weighted_fusion_4_2/mul_1
[INFO] [GpuLayer] image_resize_layer_11/ResizeNearestNeighbor
[INFO] [GpuLayer] weighted_fusion_4_2/Sum/input_Unsqueeze__2359
[INFO] [GpuLayer] weighted_fusion_3_2/strided_slice_1__963:0 + (Unnamed Layer* 787) [Shuffle] + weighted_fusion_3_2/mul_1
[INFO] [GpuLayer] weighted_fusion_3_2/Sum/input_Unsqueeze__2340
[INFO] [GpuLayer] weighted_fusion_3_2/Sum/input_Unsqueeze__2338:0 copy
[INFO] [GpuLayer] weighted_fusion_3_2/Sum/input_Unsqueeze__2340:0 copy
[INFO] [GpuLayer] weighted_fusion_3_2/Sum
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_3_2/add:0 + (Unnamed Layer* 793) [Shuffle], weighted_fusion_3_2/truediv), PWN(PWN(activation_19/Sigmoid), activation_19/mul))
[INFO] [GpuLayer] after_combine_dw_conv_3_2/separable_conv2d/depthwise__2345
[INFO] [GpuLayer] after_combine_dw_conv_3_2/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_3_2/BiasAdd
[INFO] [GpuLayer] max_pooling2d_10/MaxPool
[INFO] [GpuLayer] class-0/separable_conv2d/depthwise
[INFO] [GpuLayer] box-0/separable_conv2d/depthwise
[INFO] [GpuLayer] max_pooling2d_10/MaxPool__2355
[INFO] [GpuLayer] class-0/BiasAdd
[INFO] [GpuLayer] box-0/BiasAdd
[INFO] [GpuLayer] weighted_fusion_4_2/strided_slice_2__915:0 + (Unnamed Layer* 808) [Shuffle] + weighted_fusion_4_2/mul_2
[INFO] [GpuLayer] weighted_fusion_4_2/Sum/input_Unsqueeze__2361
[INFO] [GpuLayer] weighted_fusion_4_2/Sum/input_Unsqueeze__2357:0 copy
[INFO] [GpuLayer] weighted_fusion_4_2/Sum/input_Unsqueeze__2359:0 copy
[INFO] [GpuLayer] weighted_fusion_4_2/Sum/input_Unsqueeze__2361:0 copy
[INFO] [GpuLayer] PWN(PWN(activation_24/Sigmoid), activation_24/mul)
[INFO] [GpuLayer] PWN(PWN(activation_39/Sigmoid), activation_39/mul)
[INFO] [GpuLayer] weighted_fusion_4_2/Sum
[INFO] [GpuLayer] class-1/separable_conv2d/depthwise
[INFO] [GpuLayer] box-1/separable_conv2d/depthwise
[INFO] [GpuLayer] class-1/BiasAdd
[INFO] [GpuLayer] box-1/BiasAdd
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_4_2/add:0 + (Unnamed Layer* 822) [Shuffle], weighted_fusion_4_2/truediv), PWN(PWN(activation_20/Sigmoid), activation_20/mul))
[INFO] [GpuLayer] after_combine_dw_conv_4_2/separable_conv2d/depthwise__2366
[INFO] [GpuLayer] PWN(PWN(activation_25/Sigmoid), activation_25/mul)
[INFO] [GpuLayer] PWN(PWN(activation_40/Sigmoid), activation_40/mul)
[INFO] [GpuLayer] after_combine_dw_conv_4_2/separable_conv2d/depthwise
[INFO] [GpuLayer] class-2/separable_conv2d/depthwise
[INFO] [GpuLayer] box-2/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_4_2/BiasAdd
[INFO] [GpuLayer] class-2/BiasAdd
[INFO] [GpuLayer] box-2/BiasAdd
[INFO] [GpuLayer] max_pooling2d_11/MaxPool
[INFO] [GpuLayer] class-0_1/separable_conv2d/depthwise
[INFO] [GpuLayer] box-0_1/separable_conv2d/depthwise
[INFO] [GpuLayer] max_pooling2d_11/MaxPool__2376
[INFO] [GpuLayer] class-0_1/BiasAdd
[INFO] [GpuLayer] box-0_1/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_26/Sigmoid), activation_26/mul)
[INFO] [GpuLayer] PWN(PWN(activation_41/Sigmoid), activation_41/mul)
[INFO] [GpuLayer] weighted_fusion_5_2/strided_slice_2__867:0 + (Unnamed Layer* 855) [Shuffle] + weighted_fusion_5_2/mul_2
[INFO] [GpuLayer] class-predict/separable_conv2d/depthwise
[INFO] [GpuLayer] box-predict/separable_conv2d/depthwise
[INFO] [GpuLayer] weighted_fusion_5_2/Sum/input_Unsqueeze__2382
[INFO] [GpuLayer] class-predict/BiasAdd
[INFO] [GpuLayer] box-predict/BiasAdd
[INFO] [GpuLayer] weighted_fusion_5_2/Sum/input_Unsqueeze__2378:0 copy
[INFO] [GpuLayer] weighted_fusion_5_2/Sum/input_Unsqueeze__2380:0 copy
[INFO] [GpuLayer] weighted_fusion_5_2/Sum/input_Unsqueeze__2382:0 copy
[INFO] [GpuLayer] PWN(PWN(activation_27/Sigmoid), activation_27/mul)
[INFO] [GpuLayer] PWN(PWN(activation_42/Sigmoid), activation_42/mul)
[INFO] [GpuLayer] weighted_fusion_5_2/Sum
[INFO] [GpuLayer] class-1_1/separable_conv2d/depthwise
[INFO] [GpuLayer] box-1_1/separable_conv2d/depthwise
[INFO] [GpuLayer] class-predict/BiasAdd__2785 + Reshape
[INFO] [GpuLayer] box-predict/BiasAdd__2829 + Reshape_1
[INFO] [GpuLayer] class-1_1/BiasAdd
[INFO] [GpuLayer] box-1_1/BiasAdd
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_5_2/add:0 + (Unnamed Layer* 877) [Shuffle], weighted_fusion_5_2/truediv), PWN(PWN(activation_21/Sigmoid), activation_21/mul))
[INFO] [GpuLayer] after_combine_dw_conv_5_2/separable_conv2d/depthwise__2387
[INFO] [GpuLayer] PWN(PWN(activation_28/Sigmoid), activation_28/mul)
[INFO] [GpuLayer] PWN(PWN(activation_43/Sigmoid), activation_43/mul)
[INFO] [GpuLayer] after_combine_dw_conv_5_2/separable_conv2d/depthwise
[INFO] [GpuLayer] class-2_1/separable_conv2d/depthwise
[INFO] [GpuLayer] box-2_1/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_5_2/BiasAdd
[INFO] [GpuLayer] class-2_1/BiasAdd
[INFO] [GpuLayer] box-2_1/BiasAdd
[INFO] [GpuLayer] max_pooling2d_12/MaxPool
[INFO] [GpuLayer] class-0_2/separable_conv2d/depthwise
[INFO] [GpuLayer] box-0_2/separable_conv2d/depthwise
[INFO] [GpuLayer] max_pooling2d_12/MaxPool__2397
[INFO] [GpuLayer] class-0_2/BiasAdd
[INFO] [GpuLayer] box-0_2/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_29/Sigmoid), activation_29/mul)
[INFO] [GpuLayer] PWN(PWN(activation_44/Sigmoid), activation_44/mul)
[INFO] [GpuLayer] weighted_fusion_6_2/strided_slice_2__819:0 + (Unnamed Layer* 910) [Shuffle] + weighted_fusion_6_2/mul_2
[INFO] [GpuLayer] class-predict_1/separable_conv2d/depthwise
[INFO] [GpuLayer] box-predict_1/separable_conv2d/depthwise
[INFO] [GpuLayer] weighted_fusion_6_2/Sum/input_Unsqueeze__2403
[INFO] [GpuLayer] class-predict_1/BiasAdd
[INFO] [GpuLayer] box-predict_1/BiasAdd
[INFO] [GpuLayer] weighted_fusion_6_2/Sum/input_Unsqueeze__2399:0 copy
[INFO] [GpuLayer] weighted_fusion_6_2/Sum/input_Unsqueeze__2401:0 copy
[INFO] [GpuLayer] weighted_fusion_6_2/Sum/input_Unsqueeze__2403:0 copy
[INFO] [GpuLayer] PWN(PWN(activation_30/Sigmoid), activation_30/mul)
[INFO] [GpuLayer] PWN(PWN(activation_45/Sigmoid), activation_45/mul)
[INFO] [GpuLayer] weighted_fusion_6_2/Sum
[INFO] [GpuLayer] class-1_2/separable_conv2d/depthwise
[INFO] [GpuLayer] box-1_2/separable_conv2d/depthwise
[INFO] [GpuLayer] class-predict_1/BiasAdd__2707 + Reshape_2
[INFO] [GpuLayer] box-predict_1/BiasAdd__2746 + Reshape_3
[INFO] [GpuLayer] class-1_2/BiasAdd
[INFO] [GpuLayer] box-1_2/BiasAdd
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_6_2/add:0 + (Unnamed Layer* 932) [Shuffle], weighted_fusion_6_2/truediv), PWN(PWN(activation_22/Sigmoid), activation_22/mul))
[INFO] [GpuLayer] after_combine_dw_conv_6_2/separable_conv2d/depthwise__2408
[INFO] [GpuLayer] PWN(PWN(activation_31/Sigmoid), activation_31/mul)
[INFO] [GpuLayer] PWN(PWN(activation_46/Sigmoid), activation_46/mul)
[INFO] [GpuLayer] after_combine_dw_conv_6_2/separable_conv2d/depthwise
[INFO] [GpuLayer] class-2_2/separable_conv2d/depthwise
[INFO] [GpuLayer] box-2_2/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_6_2/BiasAdd
[INFO] [GpuLayer] class-2_2/BiasAdd
[INFO] [GpuLayer] box-2_2/BiasAdd
[INFO] [GpuLayer] max_pooling2d_13/MaxPool
[INFO] [GpuLayer] class-0_3/separable_conv2d/depthwise
[INFO] [GpuLayer] box-0_3/separable_conv2d/depthwise
[INFO] [GpuLayer] max_pooling2d_13/MaxPool__2418
[INFO] [GpuLayer] class-0_3/BiasAdd
[INFO] [GpuLayer] box-0_3/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_32/Sigmoid), activation_32/mul)
[INFO] [GpuLayer] PWN(PWN(activation_47/Sigmoid), activation_47/mul)
[INFO] [GpuLayer] weighted_fusion_7_2/strided_slice_1__786:0 + (Unnamed Layer* 965) [Shuffle] + weighted_fusion_7_2/mul_1
[INFO] [GpuLayer] class-predict_2/separable_conv2d/depthwise
[INFO] [GpuLayer] box-predict_2/separable_conv2d/depthwise
[INFO] [GpuLayer] weighted_fusion_7_2/Sum/input_Unsqueeze__2422
[INFO] [GpuLayer] class-predict_2/BiasAdd
[INFO] [GpuLayer] box-predict_2/BiasAdd
[INFO] [GpuLayer] weighted_fusion_7_2/Sum/input_Unsqueeze__2420:0 copy
[INFO] [GpuLayer] weighted_fusion_7_2/Sum/input_Unsqueeze__2422:0 copy
[INFO] [GpuLayer] PWN(PWN(activation_33/Sigmoid), activation_33/mul)
[INFO] [GpuLayer] PWN(PWN(activation_48/Sigmoid), activation_48/mul)
[INFO] [GpuLayer] weighted_fusion_7_2/Sum
[INFO] [GpuLayer] class-1_3/separable_conv2d/depthwise
[INFO] [GpuLayer] box-1_3/separable_conv2d/depthwise
[INFO] [GpuLayer] class-predict_2/BiasAdd__2629 + Reshape_4
[INFO] [GpuLayer] box-predict_2/BiasAdd__2668 + Reshape_5
[INFO] [GpuLayer] class-1_3/BiasAdd
[INFO] [GpuLayer] box-1_3/BiasAdd
[INFO] [GpuLayer] PWN(PWN(weighted_fusion_7_2/add:0 + (Unnamed Layer* 987) [Shuffle], weighted_fusion_7_2/truediv), PWN(PWN(activation_23/Sigmoid), activation_23/mul))
[INFO] [GpuLayer] after_combine_dw_conv_7_2/separable_conv2d/depthwise__2427
[INFO] [GpuLayer] PWN(PWN(activation_34/Sigmoid), activation_34/mul)
[INFO] [GpuLayer] PWN(PWN(activation_49/Sigmoid), activation_49/mul)
[INFO] [GpuLayer] after_combine_dw_conv_7_2/separable_conv2d/depthwise
[INFO] [GpuLayer] class-2_3/separable_conv2d/depthwise
[INFO] [GpuLayer] box-2_3/separable_conv2d/depthwise
[INFO] [GpuLayer] after_combine_dw_conv_7_2/BiasAdd
[INFO] [GpuLayer] class-2_3/BiasAdd
[INFO] [GpuLayer] box-2_3/BiasAdd
[INFO] [GpuLayer] class-0_4/separable_conv2d/depthwise
[INFO] [GpuLayer] box-0_4/separable_conv2d/depthwise
[INFO] [GpuLayer] class-0_4/BiasAdd
[INFO] [GpuLayer] box-0_4/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_35/Sigmoid), activation_35/mul)
[INFO] [GpuLayer] PWN(PWN(activation_50/Sigmoid), activation_50/mul)
[INFO] [GpuLayer] class-predict_3/separable_conv2d/depthwise
[INFO] [GpuLayer] box-predict_3/separable_conv2d/depthwise
[INFO] [GpuLayer] class-predict_3/BiasAdd
[INFO] [GpuLayer] box-predict_3/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_36/Sigmoid), activation_36/mul)
[INFO] [GpuLayer] PWN(PWN(activation_51/Sigmoid), activation_51/mul)
[INFO] [GpuLayer] class-1_4/separable_conv2d/depthwise
[INFO] [GpuLayer] box-1_4/separable_conv2d/depthwise
[INFO] [GpuLayer] class-predict_3/BiasAdd__2551 + Reshape_6
[INFO] [GpuLayer] box-predict_3/BiasAdd__2590 + Reshape_7
[INFO] [GpuLayer] class-1_4/BiasAdd
[INFO] [GpuLayer] box-1_4/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_37/Sigmoid), activation_37/mul)
[INFO] [GpuLayer] PWN(PWN(activation_52/Sigmoid), activation_52/mul)
[INFO] [GpuLayer] class-2_4/separable_conv2d/depthwise
[INFO] [GpuLayer] box-2_4/separable_conv2d/depthwise
[INFO] [GpuLayer] class-2_4/BiasAdd
[INFO] [GpuLayer] box-2_4/BiasAdd
[INFO] [GpuLayer] PWN(PWN(activation_38/Sigmoid), activation_38/mul)
[INFO] [GpuLayer] PWN(PWN(activation_53/Sigmoid), activation_53/mul)
[INFO] [GpuLayer] class-predict_4/separable_conv2d/depthwise
[INFO] [GpuLayer] box-predict_4/separable_conv2d/depthwise
[INFO] [GpuLayer] class-predict_4/BiasAdd
[INFO] [GpuLayer] box-predict_4/BiasAdd
[INFO] [GpuLayer] class-predict_4/BiasAdd__2473 + Reshape_8
[INFO] [GpuLayer] box-predict_4/BiasAdd__2512 + Reshape_9
[INFO] [GpuLayer] Reshape:0 copy
[INFO] [GpuLayer] Reshape_2:0 copy
[INFO] [GpuLayer] Reshape_4:0 copy
[INFO] [GpuLayer] Reshape_6:0 copy
[INFO] [GpuLayer] Reshape_8:0 copy
[INFO] [GpuLayer] Reshape_1:0 copy
[INFO] [GpuLayer] Reshape_3:0 copy
[INFO] [GpuLayer] Reshape_5:0 copy
[INFO] [GpuLayer] Reshape_7:0 copy
[INFO] [GpuLayer] Reshape_9:0 copy
[INFO] [GpuLayer] nms/anchors:0
[INFO] [GpuLayer] nms/non_maximum_suppression
[INFO] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +226, GPU +194, now: CPU 626, GPU 6035 (MiB)
[INFO] [MemUsageChange] Init cuDNN: CPU +307, GPU +310, now: CPU 933, GPU 6345 (MiB)
[WARNING] Detected invalid timing cache, setup a local cache instead
[ERROR] Tactic Device request: 2193MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2193 detected for tactic 1002.
[ERROR] Tactic Device request: 2193MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2193 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 2128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2128 detected for tactic 1002.
[ERROR] Tactic Device request: 2128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2128 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 2580MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2580 detected for tactic 1002.
[ERROR] Tactic Device request: 2580MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2580 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 1002.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 2193MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2193 detected for tactic 1002.
[ERROR] Tactic Device request: 2193MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2193 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 2128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2128 detected for tactic 1002.
[ERROR] Tactic Device request: 2128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2128 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 2580MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2580 detected for tactic 1002.
[ERROR] Tactic Device request: 2580MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2580 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 3096MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 3096 detected for tactic 1002.
[ERROR] Tactic Device request: 3096MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 3096 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 1002.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 3096MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 3096 detected for tactic 1002.
[ERROR] Tactic Device request: 3096MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 3096 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 4128 detected for tactic 20.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 4128 detected for tactic 21.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 4128 detected for tactic 22.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 4128 detected for tactic 23.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 4128 detected for tactic 24.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 4128 detected for tactic 25.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 6 due to oom error on requested size of 4128 detected for tactic 26.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 7 due to oom error on requested size of 4128 detected for tactic 27.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 8 due to oom error on requested size of 4128 detected for tactic 30.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 9 due to oom error on requested size of 4128 detected for tactic 31.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 4128 detected for tactic 128.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 4128 detected for tactic 256.
[ERROR] Tactic Device request: 4128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 4128 detected for tactic 512.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2064 detected for tactic 10.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2064 detected for tactic 11.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 2064 detected for tactic 12.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 2064 detected for tactic 13.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 2064 detected for tactic 14.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 2064 detected for tactic 15.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 6 due to oom error on requested size of 2064 detected for tactic 16.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 7 due to oom error on requested size of 2064 detected for tactic 17.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 8 due to oom error on requested size of 2064 detected for tactic 18.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 9 due to oom error on requested size of 2064 detected for tactic 19.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 10 due to oom error on requested size of 2064 detected for tactic 20.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 11 due to oom error on requested size of 2064 detected for tactic 21.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 12 due to oom error on requested size of 2064 detected for tactic 22.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 13 due to oom error on requested size of 2064 detected for tactic 23.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 14 due to oom error on requested size of 2064 detected for tactic 24.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 15 due to oom error on requested size of 2064 detected for tactic 25.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 16 due to oom error on requested size of 2064 detected for tactic 26.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 17 due to oom error on requested size of 2064 detected for tactic 27.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 18 due to oom error on requested size of 2064 detected for tactic 29.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 19 due to oom error on requested size of 2064 detected for tactic 30.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 20 due to oom error on requested size of 2064 detected for tactic 31.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2064 detected for tactic 128.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2064 detected for tactic 256.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 2064 detected for tactic 512.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 2064 detected for tactic -32.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 2064 detected for tactic -64.
[ERROR] Tactic Device request: 2064MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 2064 detected for tactic -128.
[ERROR] Tactic Device request: 2193MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2193 detected for tactic 1002.
[ERROR] Tactic Device request: 2193MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2193 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 2128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 2128 detected for tactic 1002.
[ERROR] Tactic Device request: 2128MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 2128 detected for tactic 0.
[WARNING] Internal error: cannot reformat, disabling format. Try decreasing the workspace size with IBuilderConfig::setMaxWorkspaceSize().
[WARNING] Memory requirements of format conversion cannot be satisfied during timing, format rejected.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 20.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 21.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 22.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 1548 detected for tactic 23.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 1548 detected for tactic 24.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 1548 detected for tactic 25.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 6 due to oom error on requested size of 1548 detected for tactic 26.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 7 due to oom error on requested size of 1548 detected for tactic 27.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 8 due to oom error on requested size of 1548 detected for tactic 30.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 9 due to oom error on requested size of 1548 detected for tactic 31.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 128.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 256.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 512.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 20.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 21.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 22.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 1548 detected for tactic 23.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 1548 detected for tactic 24.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 1548 detected for tactic 25.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 6 due to oom error on requested size of 1548 detected for tactic 26.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 7 due to oom error on requested size of 1548 detected for tactic 27.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 8 due to oom error on requested size of 1548 detected for tactic 30.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 9 due to oom error on requested size of 1548 detected for tactic 31.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 128.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 256.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 512.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 20.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 21.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 22.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 1548 detected for tactic 23.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 1548 detected for tactic 24.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 1548 detected for tactic 25.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 6 due to oom error on requested size of 1548 detected for tactic 26.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 7 due to oom error on requested size of 1548 detected for tactic 27.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 8 due to oom error on requested size of 1548 detected for tactic 30.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 9 due to oom error on requested size of 1548 detected for tactic 31.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 128.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 256.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 512.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 20.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 21.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 22.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 3 due to oom error on requested size of 1548 detected for tactic 23.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 4 due to oom error on requested size of 1548 detected for tactic 24.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 5 due to oom error on requested size of 1548 detected for tactic 25.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 6 due to oom error on requested size of 1548 detected for tactic 26.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 7 due to oom error on requested size of 1548 detected for tactic 27.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 8 due to oom error on requested size of 1548 detected for tactic 30.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 9 due to oom error on requested size of 1548 detected for tactic 31.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 0 due to oom error on requested size of 1548 detected for tactic 128.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 1 due to oom error on requested size of 1548 detected for tactic 256.
[ERROR] Tactic Device request: 1548MB Available: 1536MB. Device memory is insufficient to use tactic.
[WARNING] Skipping tactic 2 due to oom error on requested size of 1548 detected for tactic 512.
[INFO] Some tactics do not have sufficient workspace memory to run. Increasing workspace size may increase performance, please check verbose output.
[INFO] Detected 1 inputs and 4 output network tensors.
[INFO] Total Host Persistent Memory: 415712
[INFO] Total Device Persistent Memory: 9935360
[INFO] Total Scratch Memory: 15843328
[INFO] [MemUsageStats] Peak memory usage of TRT CPU/GPU memory allocators: CPU 12 MiB, GPU 1380 MiB
[INFO] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +0, now: CPU 1467, GPU 7146 (MiB)
[INFO] [MemUsageChange] Init cuDNN: CPU +0, GPU +0, now: CPU 1467, GPU 7146 (MiB)
[INFO] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +0, now: CPU 1467, GPU 7146 (MiB)
[INFO] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +0, now: CPU 1465, GPU 7146 (MiB)
[INFO] [MemUsageSnapshot] Builder end: CPU 1461 MiB, GPU 7146 MiB

Polygraphy Output:

$ polygraphy inspect model samples/models/test_model/grayscale_efficientdet/model.engine 
[I] Loading bytes from /opt/nvidia/deepstream/deepstream-6.0/samples/models/test_model/grayscale_efficientdet/model.engine
[I] ==== TensorRT Engine ====
    Name: Unnamed Network 0 | Explicit Batch Engine
    
    ---- 1 Engine Input(s) ----
    {image_arrays:0 [dtype=float32, shape=(1, 1376, 1024, 3)]}
    
    ---- 4 Engine Output(s) ----
    {num_detections [dtype=int32, shape=(1, 1)],
     detection_boxes [dtype=float32, shape=(1, 100, 4)],
     detection_scores [dtype=float32, shape=(1, 100)],
     detection_classes [dtype=int32, shape=(1, 100)]}
    
    ---- Memory ----
    Device Memory: 75507200 bytes
    
    ---- 1 Profile(s) (5 Binding(s) Each) ----
    - Profile: 0
        Binding Index: 0 (Input)  [Name: image_arrays:0]    | Shapes: min=(1, 1376, 1024, 3), opt=(1, 1376, 1024, 3), max=(1, 1376, 1024, 3)
        Binding Index: 1 (Output) [Name: num_detections]    | Shape: (1, 1)
        Binding Index: 2 (Output) [Name: detection_boxes]   | Shape: (1, 100, 4)
        Binding Index: 3 (Output) [Name: detection_scores]  | Shape: (1, 100)
        Binding Index: 4 (Output) [Name: detection_classes] | Shape: (1, 100)
    
    ---- 716 Layer(s) ----


May I know how did you train this grayscale model?
Could you share training spec as well? Thanks.

I’ve attached the specs file that was used for training. I did notice in this specs file that there is not an option to choose channel, so I wonder if you can only train using 3 channel?

efficientdet_d0_train.txt (1.7 KB)

After checking, the efficientdet does not support training grayscale model. So, it can only train 3 channels instead.

Thank you for checking on that! I was beginning to suspect the model could only be trained on 3 channel data. Is Nvidia planning on making an update so that EfficientDet can be trained on grayscale data?

That’s a feature request. Will let team know.
Currently, some other network can train 1 channel. For example, yolov3/yolov4/faster_rcnn/etc

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.