# NCHW and atrous in tensorflow **URL:** \ **Category:** Jetson TX2\ **Created:** [August 2, 2018, 4:36pm UTC](https://forums.developer.nvidia.com/t/nchw-and-atrous-in-tensorflow/63785 "2018-08-02T16:36:39Z")\ **Posts on this page:** 3\ **Page:** 1
**Author:** ![michael4e2ca](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@michael4e2ca](https://forums.developer.nvidia.com/u/michael4e2ca)\ **Post date:** [August 2, 2018, 4:36pm UTC](https://forums.developer.nvidia.com/t/nchw-and-atrous-in-tensorflow/63785/1 "2018-08-02T16:36:39Z")
Hi, I’m trying to convert my Tensorflow model to tensorrt to infer on tx2, I f I understand correctly Tensorrt only supports NCHW format. Tensorflow atrous\_conv2d op is only supported in NHWC so there is no way to convert with Tensorrt? Do you plan to support NHWC in the near future? Thanks ---
**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:** [August 3, 2018, 7:10am UTC](https://forums.developer.nvidia.com/t/nchw-and-atrous-in-tensorflow/63785/2 "2018-08-03T07:10:15Z")
Hi, We use NCHW implmentation for better performance. If you are using NHWC format, some conversion will be inserted when parsing the TensorFlow model. For your use-case, you may meet layer non-supported issue. A general conv2d is in our support list but not for atrous\_conv2d. Thanks. ---
**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, 6:29pm UTC](https://forums.developer.nvidia.com/t/nchw-and-atrous-in-tensorflow/63785/3 "2021-10-18T18:29:47Z")
Closed on October 18, 2021, 6:29pm UTC