While the server executable accepts a model repository path from the command line, the Server object used for tests and scripts currently hard-codes the value which works when used within the inference-server docker container using the default start up scripts but may not otherwise.
While the server executable accepts a model repository path from the command line, the Server object used for tests and scripts currently hard-codes the value which works when used within the inference-server docker container using the default start up scripts but may not otherwise.