Not supported Layer on DLA

Hi,

i’m using a Jetson AGX Xavier with Jetpack 4.3.

I’m trying to generate an tensorrt engine for one DLA core.
At the output i get following error:

Internal DLA error for layer conv1. Use allowGPUFallback() to enable GPU fallback.

The picture shows the layer conv1 with the parameters.

According to this Developer Guide :: NVIDIA Deep Learning TensorRT Documentation i don’t find the reason why the convolutional layer is not supported?

Batch size of 1 is used and the dimensions for building and runtime are the same.
Also the tensor data is less than 1 GB and fp16 is used. Furthermore the kernel dimensions, padding and stride are in the supported range.

Thanks in advance.

Hi,

The document is for our latest TensorRT version (v7.x).
But you are using TensorRT v6.0.1 from JetPack4.3.

So would you mind to upgrade your device with our latest software (JetPack4.5.1).
Then try the included TensorRT v7.1.3 instead?

Thanks.

Thanks for your quick response.

Currently it is not possible to upgrade.
Are there any documentations for TensorRT v6.0.1?

Hi,

You can find all the document in the Archives:
https://docs.nvidia.com/deeplearning/tensorrt/archives/index.html

Thanks.