# SSD model to TensorRT with customer layer

**URL:** <https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783>\
**Category:** Jetson TX2\
**Created:** [March 5, 2018, 5:13am UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783 "2018-03-05T05:13:17Z")\
**Posts on this page:** 7\
**Page:** 1

<div class="post-metadata">

**Author:** ![hi\_fly](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@hi\_fly](https://forums.developer.nvidia.com/u/hi_fly)\
**Post date:** [March 5, 2018, 5:13am UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/1 "2018-03-05T05:13:17Z")

</div>

Hi,I am trying to convert SSD model(VGG\_VOC0712\_SSD\_300x300\_deploy) to run with TensorRT(TX2) use customer layer.

I have my sample compiled successfully but get an error when running:

sample\_SSD: NvPluginSSD.cu:587: virtual void nvinfer1::plugin::DetectionOutput::configure(const nvinfer1::Dims\*, int, const nvinfer1::Dims\*, int, int): Assertion `numPriors_numLocClasses_4 == inputDims[0].d[0]’ failed.

Thank you so much for any advise and help.

---

<div class="post-metadata">

**Author:** ![hi\_fly](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@hi\_fly](https://forums.developer.nvidia.com/u/hi_fly)\
**Post date:** [March 5, 2018, 6:52am UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/2 "2018-03-05T06:52:59Z")

</div>

My mistake ,it seems like some parameters of the layer was wrong.

---

<div class="post-metadata">

**Author:** ![AastaLLL](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/aastalll/32/14043_2.png) [@AastaLLL](https://forums.developer.nvidia.com/u/AastaLLL)\
**Post date:** [March 5, 2018, 6:55am UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/3 "2018-03-05T06:55:58Z")

</div>

Hi,

SSD plugins are not official release functions.  
We don’t have information can share with you.

Here is a relevant discussion for your reference:

> <https://github.com/dusty-nv/jetson-inference/issues/171>
>
> I'm trying to build the objection detection framework SSD with TensorRT 3.0 plug…in api, the DetectionOutput layer was created by the create function in the NVIDIA plugin library createSSDDetectionOutputPlugin. But it does return a error message "Plugin layer output count is not equal to caffe output count".When I remove the DetectionOutput layer in the prototxt file, it can run and I can get the output of the mbox\_loc, mbox\_conf\_flatten, mbox\_priorbox layer.
> Can you help me ? Thank you!

Thanks.

---

<div class="post-metadata">

**Author:** ![xxdyx110](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@xxdyx110](https://forums.developer.nvidia.com/u/xxdyx110)\
**Post date:** [July 13, 2018, 12:17pm UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/4 "2018-07-13T12:17:03Z")

</div>

I met the same error, have u solve it?  
I use it in voc\_ssd, it’s ok. But when it come to mobilenet\_ssd, the error occurs.  
hope your reply~

---

<div class="post-metadata">

**Author:** ![AastaLLL](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/aastalll/32/14043_2.png) [@AastaLLL](https://forums.developer.nvidia.com/u/AastaLLL)\
**Post date:** [July 23, 2018, 6:51am UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/5 "2018-07-23T06:51:41Z")

</div>

Hi,

Do you use Caffe model and want to inference it with our internal SSD plugin function (ex, createSSDNormalizePlugin,…)?  
Please remember that these SSD plugin implementation is not official supported.  
It’s recommended to use our new sampleUffSSD sample in TensorRT 4.0 instead.

If you are not using the internal SSD plugin function, could you file a new topic specify for your issue?

Thanks.

---

<div class="post-metadata">

**Author:** ![285372433](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/285372433/32/11468_2.png) [@285372433](https://forums.developer.nvidia.com/u/285372433)\
**Post date:** [October 10, 2018, 8:05am UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/6 "2018-10-10T08:05:01Z")

</div>

> [@](#):
>
> I met the same error, have u solve it?  
> I use it in voc\_ssd, it’s ok. But when it come to mobilenet\_ssd, the error occurs.  
> hope your reply~

Hi  
Did you solve this problem? I have the same error Assertion failed.

---

<div class="post-metadata">

**Author:** ![nadeemm](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/nadeemm/32/367020_2.png) [@nadeemm](https://forums.developer.nvidia.com/u/nadeemm)\
**Post date:** [October 18, 2021, 6:30pm UTC](https://forums.developer.nvidia.com/t/ssd-model-to-tensorrt-with-customer-layer/58783/7 "2021-10-18T18:30:27Z")

</div>

Closed on October 18, 2021, 6:30pm UTC


