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/providers/remote/post_training/nvidia/README.md

@@ -36,7 +36,6 @@ import os
 
 os.environ["NVIDIA_API_KEY"] = "your-api-key"
 os.environ["NVIDIA_CUSTOMIZER_URL"] = "http://nemo.test"
-os.environ["NVIDIA_USER_ID"] = "llama-stack-user"
 os.environ["NVIDIA_DATASET_NAMESPACE"] = "default"
 os.environ["NVIDIA_PROJECT_ID"] = "test-project"
 os.environ["NVIDIA_OUTPUT_MODEL_DIR"] = "test-example-model@v1"
@@ -125,6 +124,21 @@ client.post_training.job.cancel(job_uuid="your-job-id")
 
 ### Inference with the fine-tuned model
 
+#### 1. Register the model
+
+```python
+from llama_stack.apis.models import Model, ModelType
+
+client.models.register(
+    model_id="test-example-model@v1",
+    provider_id="nvidia",
+    provider_model_id="test-example-model@v1",
+    model_type=ModelType.llm,
+)
+```
+
+#### 2. Inference with the fine-tuned model
+
 ```python
 response = client.inference.completion(
     content="Complete the sentence using one word: Roses are red, violets are ",