Dockerized version of tf-lift from https://github.com/cvlab-epfl/tf-lift
864
A docker build for tf-lift. When running you need to use the runtime nvidia to pass the driver to the docker. The pretrained models are also included.
When running the docker you will end up in the tf-lift folder. From here you can run the examples as in the original project. For example, saving the keypoints of someimage.jpg to someimage.txt using the pretrained kp network:
docker pull fremmen/tf-lift
docker run -it --rm --runtime=nvidia fremmen/tf-lift
$ ~/tf-lift: python main.py --task=test --subtask=kp --logdir=release-aug --test_img_file=someimage.jpg --test_out_file=someimage.txt --use_batch_norm=false --mean_std_type=hardcoded
Content type
Image
Digest
Size
1.3 GB
Last updated
about 7 years ago
docker pull fremmen/tf-lift