New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
xgboost models will not load #309
Comments
Default PyClosureContainerDockerfile and python_container_conda_deps.txt don't have the xgboost module installed. You must create own DockerFile to install xgboost module according to the this reference. |
Yep, I did install xgboost using a separate Dockerfile and called it as the build image. I confirmed this after running docker exec pip list and xgboost is shown as installed. So I don’t believe this is an import error. |
Your error log shows that Clipper failed to call LoadLibrary() in your own container. I think that xgboost module needs to reinstall, maybe to rebuild. (refer to this link: dmlc/xgboost#463) |
Ok I figured it out. Docker requires the right gcc compiler in order for xgboost to work. However, even after installing gcc, it still doesn't compile correctly. That's because the default Clipper conda environ has an outdated version of This is my Dockerfile that I got to run:
|
Hello,
I've trying to use Clipper to deploy an xgboost model. However, the model container instantly exits after I run the python deployer. This is the traceback I'm getting from the logs:
I'm not sure if this is an issue with Clipper or with the compiler within the docker container. I don't know why the model container would exit immediately on deployment like this?
The text was updated successfully, but these errors were encountered: