Open ezar opened 5 years ago
GLBurnett
commented
5 years ago Hey @ezar,
The object_detection_webcam container is built on top of a TensorRT container from the NVIDIA GPU Cloud (NGC). The container shouldn't require a log in, but since you are getting this error, I would try making an account on NGC, which contains an NVIDIA run Docker repository, and then doing a docker login, using your NGC credentials.
More detailed instructions can be found on the NGC website.
Dask91
commented
4 years ago I concur, unfortunately the Dockerfile does not appear to work. I don't believe this issue pertains to the docker login, as I am already logged in the NGC. The problem rather seems to be with the RUN commands.
mckelvaney
commented
4 years ago @ezar @Dask91 are you trying to run this on an arm architecture device (Jetson Nano or similar)? It would appear that this image is built on x86, so will not run on arm. The message exec format error generally indicates this.
Tested by successfully running the image on MacOS 10.14 and unsuccessfully running the image on Jetson Nano.
standard_init_linux.go:207: exec user process caused "exec format error" The command '/bin/sh -c apt-get update && apt-get install -y --no-install-recommends libcurl4-openssl-dev wget zlib1g-dev git pkg-config python3 python3-pip' returned a non-zero code: 1 Starting docker container Unable to find image 'object_detection_webcam:latest' locally docker: Error response from daemon: pull access denied for object_detection_webcam, repository does not exist or may require 'docker login'.