Convert SC-FEGAN checkpoint to TensorRT


I have been using SC-FEGAN model for facial parts inpainting using Tensorflow 1.13 and it works with no issues on my desktop.
The issue is when I try to run inference on my jetson nano 4G (I got OOM out of memory)
So I thought about converting the SC-FEGAN used checkpoint to TensorRt and re-test.
The issue is that the provided files are only : and SC-FEGAN.ckpt.index

So is there any possible way to convert the given Checkpoint to TensorRT since I miss the .meta file?


TensorRT Version: 5.0
GPU Type: GTX 1650 Ti
Nvidia Driver Version: 460.32
CUDA Version: 10.0
CUDNN Version: 7.6.2
Operating System + Version: Ubuntu 18.04.6
Python Version (if applicable): 3.7
TensorFlow Version (if applicable): 1.13.1

Relevant Files

Github repo: GitHub - run-youngjoo/SC-FEGAN: SC-FEGAN : Face Editing Generative Adversarial Network with User's Sketch and Color (ICCV2019)
Google Drive (Checkpoints): FC-FEGAN_ckpt - Google Drive

We recommend you to check the below samples links in case of tf-trt integration issues.

If issue persist, We recommend you to reach out to Tensorflow forum.

1 Like