add NVIDIA_BASE_URL and NVIDIA_API_KEY to control hosted vs local endpoints

[mattf]

What does this PR do?

allows template distribution connect to hosted or local NIM:

use --env NVIDIA_BASE_URL=http://localhost:8000 to connect to a local NIM running at localhost:8000

use --env NVIDIA_API_KEY=blah when connecting to hosted NIM, e.g. NVIDIA_BASE_URL=https://integrate.api.nvidia.com

Test Plan

• llama stack run ./llama_stack/templates/nvidia/run.yaml -> error, e.g. API key is required for hosted NVIDIA NIM

• llama stack run ./llama_stack/templates/nvidia/run.yaml --env NVIDIA_BASE_URL=https://integrate.api.nvidia.com -> error, e.g. API key is required for hosted NVIDIA NIM

• llama stack run ./llama_stack/templates/nvidia/run.yaml --env NVIDIA_API_KEY=REDACTED -> successful connection to NIM on https://integrate.api.nvidia.com

• llama stack run ./llama_stack/templates/nvidia/run.yaml --env NVIDIA_BASE_URL=https://integrate.api.nvidia.com --env NVIDIA_API_KEY=REDACTED -> successful connection to NIM running on integrate.api.nvidia.com

• llama stack run ./llama_stack/templates/nvidia/run.yaml --env NVIDIA_BASE_URL=http://localhost:8000 -> successful connection to NIM running on localhost:8000

• llama stack run ./llama_stack/templates/nvidia/run.yaml --env NVIDIA_BASE_URL=http://localhost:8000 --env NVIDIA_API_KEY=REDACTED -> successful connection to NIM running on http://localhost:8000

• llama stack run ./llama_stack/templates/nvidia/run.yaml --env NVIDIA_BASE_URL=http://bogus -> runtime error, e.g. ConnectionError (TODO: this should be a startup error)

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Ran pre-commit to handle lint / formatting issues.
• Read the contributor guideline,
  Pull Request section?
• Updated relevant documentation.
• Wrote necessary unit or integration tests.

[mattf]

cc @cdgamarose-nv

[ashwinb]

thanks