failed to load the pre training model

Question

failed to load the pre training model

Closed this issue 2 years ago · 1 comments

Hello, failed to load the pre training model. The environment is cuda11.4. What's the reason

terminate called after throwing an instance of 'std::runtime_error'
what(): Failed importing graph from: data/nets/scannet256_32_frozen.pb

Answer 1 · 2022-11-18T09:27:55.000Z

Unfortunately I am no longer maintaining this project due to restrictions of my current employment. Last time I tested the software was in our docker image, which uses nvidia/cuda:10.0-cudnn7-devel as the base image. I suppose you will have the best luck running the software if you switch to the CUDA and cudnn version from that image, or simply build and use the docker image.