feat: NVIDIA allow non-llama model registration (#1859)

# What does this PR do?
Adds custom model registration functionality to NVIDIAInferenceAdapter
which let's the inference happen on:
- post-training model
- non-llama models in API Catalogue(behind
https://integrate.api.nvidia.com and endpoints compatible with
AyncOpenAI)

## Example Usage:
```python
from llama_stack.apis.models import Model, ModelType
from llama_stack.distribution.library_client import LlamaStackAsLibraryClient
client = LlamaStackAsLibraryClient("nvidia")
_ = client.initialize()

client.models.register(
        model_id=model_name,
        model_type=ModelType.llm,
        provider_id="nvidia"
)

response = client.inference.chat_completion(
    model_id=model_name,
    messages=[{"role":"system","content":"You are a helpful assistant."},{"role":"user","content":"Write a limerick about the wonders of GPU computing."}],
)
```

## Test Plan
```bash
pytest tests/unit/providers/nvidia/test_supervised_fine_tuning.py 
========================================================== test session starts ===========================================================
platform linux -- Python 3.10.0, pytest-8.3.5, pluggy-1.5.0
rootdir: /home/ubuntu/llama-stack
configfile: pyproject.toml
plugins: anyio-4.9.0
collected 6 items                                                                                                                        

tests/unit/providers/nvidia/test_supervised_fine_tuning.py ......                                                                  [100%]

============================================================ warnings summary ============================================================
../miniconda/envs/nvidia-1/lib/python3.10/site-packages/pydantic/fields.py:1076
  /home/ubuntu/miniconda/envs/nvidia-1/lib/python3.10/site-packages/pydantic/fields.py:1076: PydanticDeprecatedSince20: Using extra keyword arguments on `Field` is deprecated and will be removed. Use `json_schema_extra` instead. (Extra keys: 'contentEncoding'). Deprecated in Pydantic V2.0 to be removed in V3.0. See Pydantic V2 Migration Guide at https://errors.pydantic.dev/2.11/migration/
    warn(

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
====================================================== 6 passed, 1 warning in 1.51s ======================================================
```

[//]: # (## Documentation)
Updated Readme.md

cc: @dglogo, @sumitb, @mattf

llama_stack/templates/nvidia/nvidia.py

@@ -98,19 +98,15 @@ def get_distribution_template() -> DistributionTemplate:
                 "",
                 "NVIDIA API Key",
             ),
-            ## Nemo Customizer related variables
-            "NVIDIA_USER_ID": (
-                "llama-stack-user",
-                "NVIDIA User ID",
+            "NVIDIA_APPEND_API_VERSION": (
+                "True",
+                "Whether to append the API version to the base_url",
             ),
+            ## Nemo Customizer related variables
             "NVIDIA_DATASET_NAMESPACE": (
                 "default",
                 "NVIDIA Dataset Namespace",
             ),
-            "NVIDIA_ACCESS_POLICIES": (
-                "{}",
-                "NVIDIA Access Policies",
-            ),
             "NVIDIA_PROJECT_ID": (
                 "test-project",
                 "NVIDIA Project ID",

llama_stack/templates/nvidia/run-with-safety.yaml

@@ -18,6 +18,7 @@ providers:
     config:
       url: ${env.NVIDIA_BASE_URL:https://integrate.api.nvidia.com}
       api_key: ${env.NVIDIA_API_KEY:}
+      append_api_version: ${env.NVIDIA_APPEND_API_VERSION:True}
   - provider_id: nvidia
     provider_type: remote::nvidia
     config:

llama_stack/templates/nvidia/run.yaml

@@ -18,6 +18,7 @@ providers:
     config:
       url: ${env.NVIDIA_BASE_URL:https://integrate.api.nvidia.com}
       api_key: ${env.NVIDIA_API_KEY:}
+      append_api_version: ${env.NVIDIA_APPEND_API_VERSION:True}
   vector_io:
   - provider_id: faiss
     provider_type: inline::faiss