feat: Add NVIDIA Eval integration (#1890)

# What does this PR do?
This PR adds support for NVIDIA's NeMo Evaluator API to the Llama Stack
eval module. The integration enables users to evaluate models via the
Llama Stack interface.

## Test Plan
[Describe the tests you ran to verify your changes with result
summaries. *Provide clear instructions so the plan can be easily
re-executed.*]
1. Added unit tests and successfully ran from root of project:
`./scripts/unit-tests.sh tests/unit/providers/nvidia/test_eval.py`
```
tests/unit/providers/nvidia/test_eval.py::TestNVIDIAEvalImpl::test_job_cancel PASSED
tests/unit/providers/nvidia/test_eval.py::TestNVIDIAEvalImpl::test_job_result PASSED
tests/unit/providers/nvidia/test_eval.py::TestNVIDIAEvalImpl::test_job_status PASSED
tests/unit/providers/nvidia/test_eval.py::TestNVIDIAEvalImpl::test_register_benchmark PASSED
tests/unit/providers/nvidia/test_eval.py::TestNVIDIAEvalImpl::test_run_eval PASSED
```
2. Verified I could build the Llama Stack image: `LLAMA_STACK_DIR=$(pwd)
llama stack build --template nvidia --image-type venv`

Documentation added to
`llama_stack/providers/remote/eval/nvidia/README.md`

---------

Co-authored-by: Jash Gulabrai <jgulabrai@nvidia.com>

llama_stack/templates/dependencies.json

@@ -394,12 +394,10 @@
     "aiosqlite",
     "blobfile",
     "chardet",
-    "emoji",
     "faiss-cpu",
     "fastapi",
     "fire",
     "httpx",
-    "langdetect",
     "matplotlib",
     "nltk",
     "numpy",
@@ -411,7 +409,6 @@
     "psycopg2-binary",
     "pymongo",
     "pypdf",
-    "pythainlp",
     "redis",
     "requests",
     "scikit-learn",
@@ -419,7 +416,6 @@
     "sentencepiece",
     "tqdm",
     "transformers",
-    "tree_sitter",
     "uvicorn"
   ],
   "ollama": [

llama_stack/templates/nvidia/build.yaml

@@ -1,6 +1,6 @@
 version: '2'
 distribution_spec:
-  description: Use NVIDIA NIM for running LLM inference and safety
+  description: Use NVIDIA NIM for running LLM inference, evaluation and safety
   providers:
     inference:
     - remote::nvidia
@@ -13,7 +13,7 @@ distribution_spec:
     telemetry:
     - inline::meta-reference
     eval:
-    - inline::meta-reference
+    - remote::nvidia
     post_training:
     - remote::nvidia
     datasetio:

llama_stack/templates/nvidia/nvidia.py

@@ -7,6 +7,7 @@
 from pathlib import Path
 
 from llama_stack.distribution.datatypes import ModelInput, Provider, ShieldInput, ToolGroupInput
+from llama_stack.providers.remote.eval.nvidia import NVIDIAEvalConfig
 from llama_stack.providers.remote.inference.nvidia import NVIDIAConfig
 from llama_stack.providers.remote.inference.nvidia.models import MODEL_ENTRIES
 from llama_stack.providers.remote.safety.nvidia import NVIDIASafetyConfig
@@ -20,7 +21,7 @@ def get_distribution_template() -> DistributionTemplate:
         "safety": ["remote::nvidia"],
         "agents": ["inline::meta-reference"],
         "telemetry": ["inline::meta-reference"],
-        "eval": ["inline::meta-reference"],
+        "eval": ["remote::nvidia"],
         "post_training": ["remote::nvidia"],
         "datasetio": ["inline::localfs"],
         "scoring": ["inline::basic"],
@@ -37,6 +38,11 @@ def get_distribution_template() -> DistributionTemplate:
         provider_type="remote::nvidia",
         config=NVIDIASafetyConfig.sample_run_config(),
     )
+    eval_provider = Provider(
+        provider_id="nvidia",
+        provider_type="remote::nvidia",
+        config=NVIDIAEvalConfig.sample_run_config(),
+    )
     inference_model = ModelInput(
         model_id="${env.INFERENCE_MODEL}",
         provider_id="nvidia",
@@ -60,7 +66,7 @@ def get_distribution_template() -> DistributionTemplate:
     return DistributionTemplate(
         name="nvidia",
         distro_type="self_hosted",
-        description="Use NVIDIA NIM for running LLM inference and safety",
+        description="Use NVIDIA NIM for running LLM inference, evaluation and safety",
         container_image=None,
         template_path=Path(__file__).parent / "doc_template.md",
         providers=providers,
@@ -69,6 +75,7 @@ def get_distribution_template() -> DistributionTemplate:
             "run.yaml": RunConfigSettings(
                 provider_overrides={
                     "inference": [inference_provider],
+                    "eval": [eval_provider],
                 },
                 default_models=default_models,
                 default_tool_groups=default_tool_groups,
@@ -78,7 +85,8 @@ def get_distribution_template() -> DistributionTemplate:
                     "inference": [
                         inference_provider,
                         safety_provider,
-                    ]
+                    ],
+                    "eval": [eval_provider],
                 },
                 default_models=[inference_model, safety_model],
                 default_shields=[ShieldInput(shield_id="${env.SAFETY_MODEL}", provider_id="nvidia")],
@@ -119,6 +127,10 @@ def get_distribution_template() -> DistributionTemplate:
                 "http://0.0.0.0:7331",
                 "URL for the NeMo Guardrails Service",
             ),
+            "NVIDIA_EVALUATOR_URL": (
+                "http://0.0.0.0:7331",
+                "URL for the NeMo Evaluator Service",
+            ),
             "INFERENCE_MODEL": (
                 "Llama3.1-8B-Instruct",
                 "Inference model",

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

@@ -53,13 +53,10 @@ providers:
       sinks: ${env.TELEMETRY_SINKS:console,sqlite}
       sqlite_db_path: ${env.SQLITE_DB_PATH:~/.llama/distributions/nvidia/trace_store.db}
   eval:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
+  - provider_id: nvidia
+    provider_type: remote::nvidia
     config:
-      kvstore:
-        type: sqlite
-        namespace: null
-        db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/meta_reference_eval.db
+      evaluator_url: ${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}
   post_training:
   - provider_id: nvidia
     provider_type: remote::nvidia

llama_stack/templates/nvidia/run.yaml

@@ -48,13 +48,10 @@ providers:
       sinks: ${env.TELEMETRY_SINKS:console,sqlite}
       sqlite_db_path: ${env.SQLITE_DB_PATH:~/.llama/distributions/nvidia/trace_store.db}
   eval:
-  - provider_id: meta-reference
-    provider_type: inline::meta-reference
+  - provider_id: nvidia
+    provider_type: remote::nvidia
     config:
-      kvstore:
-        type: sqlite
-        namespace: null
-        db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/meta_reference_eval.db
+      evaluator_url: ${env.NVIDIA_EVALUATOR_URL:http://localhost:7331}
   post_training:
   - provider_id: nvidia
     provider_type: remote::nvidia