EDSR model(Superscaling) implimentation Steps

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

Description

Hi , I have a tensorflow edsr model and I have convered the same to onnx and then to a .trt engine file
Kindly tell how to get inference and whether any samples for the same is present

Environment

TensorRT Version: 7.0
GPU Type: dGPU T4
Nvidia Driver Version:440.9
CUDA Version: 10.2
Operating System + Version:18.04
Python Version (if applicable): 2.7 and 3.6

2

Hi @GalibaSashi,
Please take reference from below example.

Thanks!

3

Hi @AakankshaS
Actually here upscaling is needed, So detection classification and segmentation wont work.

4

Hi @GalibaSashi,
As per my knowledge we don’t have any such open source sample.
These are the list of samples available with python implementation from where i have shared the previous link.

You can try trtexec for bench-marking and debugging your model.
https://github.com/NVIDIA/TensorRT/tree/master/samples/opensource/trtexec

Thanks!

5

I can use sudo ./trtexec --onnx=/home/bhmk/Desktop/ML_BHMK/USECASE3/Models/edsr.onnx --verbose --explicitBatch --shapes=inputShapes --saveEngine=edsr.trt successfully.
Can you tell the command to benchmark the same model.

6

Hi @GalibaSashi,
You will get all implementations of trtexec on the link.
Benchmarking sample is in the link below
https://github.com/NVIDIA/TensorRT/tree/master/samples/opensource/trtexec#example-1-simple-mnist-model-from-caffe

Thanks!

7

Actually I wanted to test it on a image given by me itself. CAn you tell me the steps.

8

@GalibaSashi did you manage to get inference with the edsr model with trt engine format?