Update model paths to use llm_models_root instead of HF

Signed-off-by: Gal Hubara-Agam <96368689+galagam@users.noreply.github.com>

Author: galagam

tests/integration/defs/accuracy/test_llm_api_autodeploy.py

@@ -264,9 +264,9 @@ class TestNemotronSuperV3(LlmapiAccuracyTestHarness):
     """
 
     MODEL_NAME = "nvidia/NVIDIA-Nemotron-3-Super-120B-012726"
-    MODEL_PATH_BF16 = f"nvidia/NVIDIA-Nemotron-3-Super-120B-BF16-BF16KV-012726"
-    MODEL_PATH_FP8 = "nvidia/NVIDIA-Nemotron-3-Super-120B-FP8-FP8KV-012726"
-    MODEL_PATH_FP4 = f"nvidia/NVIDIA-Nemotron-3-Super-120B-NVFP4-FP8KV-012726"
+    MODEL_PATH_BF16 = f"{llm_models_root()}/NVIDIA-Nemotron-3-Super-120B-BF16-BF16KV-012726"
+    MODEL_PATH_FP8 = f"{llm_models_root()}/NVIDIA-Nemotron-3-Super-120B-FP8-FP8KV-012726"
+    MODEL_PATH_FP4 = f"{llm_models_root()}/NVIDIA-Nemotron-3-Super-120B-NVFP4-FP8KV-012726"
 
     # Set minimum possible seq len + small buffer, for test speed & memory usage
     MAX_SEQ_LEN = max(MMLU.MAX_INPUT_LEN + MMLU.MAX_OUTPUT_LEN,