# Custom Layer Tensorflow support for TensorRT 3

**URL:** <https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016>\
**Category:** General\
**Created:** [February 8, 2018, 12:51pm UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016 "2018-02-08T12:51:20Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![krunal.vaghani](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@krunal.vaghani](https://forums.developer.nvidia.com/u/krunal.vaghani)\
**Post date:** [February 8, 2018, 12:51pm UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016/1 "2018-02-08T12:51:20Z")

</div>

Is TensorRT provides a Custom Layer C++ API for Tensorflow to inference UFF file on DrivePX2?

---

<div class="post-metadata">

**Author:** ![SteveNV](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/stevenv/32/14043_2.png) [@SteveNV](https://forums.developer.nvidia.com/u/SteveNV)\
**Post date:** [February 9, 2018, 8:16pm UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016/2 "2018-02-09T20:16:04Z")

</div>

Dear krunal.vaghani,

Unfortunately, if use any network that requires custom layer in TensorRT, cannot use DriveWorks API for DNNs.

---

<div class="post-metadata">

**Author:** ![krunal.vaghani](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@krunal.vaghani](https://forums.developer.nvidia.com/u/krunal.vaghani)\
**Post date:** [February 12, 2018, 6:58am UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016/3 "2018-02-12T06:58:46Z")

</div>

Hey,

Okay!

One mode thing, In TensorRT sample list, there is one samplePlugin example which shows that we can implement a custom caffe layer using Plugin API of TensorRT using NvCaffeParser.

So, you are saying that, if I use this custom layer DNN in TensorRT for example pedestrian detection network and that will not support on DrivePX2?

As for custom layered caffe networks, TensorRT has Plugin API so, for custom layered Tensorflow Model there is no way I mean like Plugin API that we can use TensorFlow model to inference on drivePX2?

---

<div class="post-metadata">

**Author:** ![maximilian.fink](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@maximilian.fink](https://forums.developer.nvidia.com/u/maximilian.fink)\
**Post date:** [May 4, 2018, 5:45am UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016/4 "2018-05-04T05:45:20Z")

</div>

Hi,

so I would like to have a clear statement from you SteveNV. Is it possible to use a Tensorflow model with the UFF parser together with custom layers and do the integration on the DRIVE PX2? Yes or No? If yes, are there any examples? If no, is it planned in the near future to support plugins for Tensorflow models or do we have to switch to Caffe framework?

Thanks.

---

<div class="post-metadata">

**Author:** ![tudor626a1](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@tudor626a1](https://forums.developer.nvidia.com/u/tudor626a1)\
**Post date:** [July 27, 2018, 11:12pm UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016/5 "2018-07-27T23:12:11Z")

</div>

We are also wondering the same thing.

---

<div class="post-metadata">

**Author:** ![SteveNV](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/stevenv/32/14043_2.png) [@SteveNV](https://forums.developer.nvidia.com/u/SteveNV)\
**Post date:** [July 30, 2018, 2:15am UTC](https://forums.developer.nvidia.com/t/custom-layer-tensorflow-support-for-tensorrt-3/58016/6 "2018-07-30T02:15:30Z")

</div>

Dear All,

Could you please refer to the following link for your topic? Thanks.

[https://devtalk.nvidia.com/default/topic/1030068/driveworks/-solved-tensorrt3-tensorflow-implementation-in-px2/](https://devtalk.nvidia.com/default/topic/1030068/driveworks/-solved-tensorrt3-tensorflow-implementation-in-px2/)