Hello,
I am experiencing an issue when trying to start the chat feature on NVIDIA Workbench. The process keeps polling the inference server without success. Below is the error log I am encountering:
I have reviewed the provided postbuild script in my project, which sets up the required environments and dependencies. The script appears to complete successfully, but the issue persists when starting the chat.
hi. is this a project you build or one of the examples?
I meet this issue also.
Hi @Alex.Y sorry you’ve run into this. Is this on a project that you have built, or one of the example workbench projects?
example project. And I call help to Nvidia, their engineer kindly help privde a way to fix this with adding the variable /usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH and add “$HOME/.conda/envs/api-env/bin/pip install --upgrade setuptools” to the postBuild.bash
Pleased you were able to get this resolved. Thanks for sharing the solution here for others too - have a great weekend!
