Issue while trying with ssd_inception_v2_coco from tensorflow zoo about tensorrt_demos

[jkjung-avt]

I think the issue could be fixed by just removing the "image_tensor:0" input by enabling this line of code: https://github.com/jkjung-avt/tensorrt_demos/blob/master/ssd/build_engine.py#L160

[SahilChachere]

I think the issue could be fixed by just removing the "image_tensor:0" input by enabling this line of code: https://github.com/jkjung-avt/tensorrt_demos/blob/master/ssd/build_engine.py#L160

I solved this but now at last step during conversion I am getting core dumped error.

for reference:

UFF Output written to /home/sys-admin/Downloads/Projects/tensorrt_demos/ssd/tmp_inception_v2_coco.uff

UFF Text Output written to /home/sys-admin/Downloads/Projects/tensorrt_demos/ssd/tmp_inception_v2_coco.pbtxt

[TensorRT] INFO: Detected 1 inputs and 2 output network tensors.
python3: nmsPlugin.cpp:139: virtual void nvinfer1::plugin::DetectionOutput::configureWithFormat(const nvinfer1::Dims*, int, const nvinfer1::Dims*, int, nvinfer1::DataType, nvinfer1::PluginFormat, int): Assertion `numPriors * numLocClasses * 4 == inputDims[param.inputOrder[0]].d[0]' failed.
./build_engines.sh: line 5: 7298 Aborted (core dumped) python3 build_engine.py ${model}

[jkjung-avt]

Check input order of the 3 input tensors of the NMS_TRT node (look at your 'tmp_inception_v2_coco.pbtxt').

Reference: https://devtalk.nvidia.com/default/topic/1050465/jetson-nano/how-to-write-config-py-for-converting-ssd-mobilenetv2-to-uff-format/post/5333033/#5333033

[SahilChachere]

Check input order of the 3 input tensors of the NMS_TRT node (look at your 'tmp_inception_v2_coco.pbtxt').

Reference: https://devtalk.nvidia.com/default/topic/1050465/jetson-nano/how-to-write-config-py-for-converting-ssd-mobilenetv2-to-uff-format/post/5333033/#5333033

@jkjung-avt as you mentioned to check 'tmp_inception_v2_coco.pbtxt' file, I have gone through it and compared parameters with build_engine.py file, all are similar, still, I am getting core dumped error, have you tried with ssd_inception model or have you such issue, what solution can you suggest specifically for ssd_inception_v2 model?

for reference:
in build_engine.py file -

},
    'ssd_inception_v2_coco': {
        'input_pb':   os.path.abspath(os.path.join(
                          DIR_NAME, 'ssd_inception_v2_coco.pb')),
        'tmp_uff':    os.path.abspath(os.path.join(
                          DIR_NAME, 'tmp_inception_v2_coco.uff')),
        'output_bin': os.path.abspath(os.path.join(
                          DIR_NAME, 'TRT_ssd_inception_v2_coco.bin')),
        'num_classes': 90,
        'min_size': 0.2,
        'max_size': 0.95,
        'input_order': [1, 0, 2],  # order of loc_data, conf_data, priorbox_data
    },

what else changes I must do in code??

[SahilChachere]

@jkjung-avt thanks for replies, I solved all issues and successfully converted & run by changing 'input_order': [1, 0, 2], to 'input_order': [0, 2, 1],