[Serve] Getting Started doc: Converting to a Ray Serve deployment: Various issues

[angelinalg]

Description

1. The code example in the Converting-to-a-ray-serve-deployment section has line numbers. If you use the copy and paste button, it includes the line numbers and the user needs to remove them. Is there a particular reason for including line numbers for this example?

2. Suggest spelling out the python serve_deployment.py command.

3. Getting these errors: (ServeReplica:Translator pid=13568) 2023-04-17 11:42:32.206766: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/nvidia/lib:/usr/local/nvidia/lib64 (ServeReplica:Translator pid=13568) 2023-04-17 11:42:32.206800: W tensorflow/stream_executor/cuda/cuda_driver.cc:269] failed call to cuInit: UNKNOWN ERROR (303) (ServeReplica:Translator pid=13567) 2023-04-17 11:42:32.207429: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/nvidia/lib:/usr/local/nvidia/lib64 (ServeReplica:Translator pid=13567) 2023-04-17 11:42:32.207464: W tensorflow/stream_executor/cuda/cuda_driver.cc:269] failed call to cuInit: UNKNOWN ERROR (303)

4. Add guidance on what to expect after the last step. Should the reader just get the message that says, Deployed Serve app successfully?

5. Is this a typo? python model.py -> python summary_model.py

Link

https://docs.ray.io/en/latest/serve/getting_started.html#converting-to-a-ray-serve-deployment

[angelinalg]

cc: @akshay-anyscale

[angelinalg]

This PR (34617) fixes #1 and #2 but not the rest.