rename impl, add config validation

llama_stack/providers/remote/post_training/nvidia/__init__.py

@@ -10,14 +10,18 @@ from llama_stack.distribution.datatypes import Api, ProviderSpec
 
 from .config import NvidiaPostTrainingConfig
 
-# post_training api and the torchtune provider is still experimental and under heavy development
-
 
 async def get_adapter_impl(
     config: NvidiaPostTrainingConfig,
     deps: Dict[Api, ProviderSpec],
 ):
-    from .post_training import NvidiaPostTrainingImpl
+    from .post_training import NvidiaPostTrainingAdapter
 
-    impl = NvidiaPostTrainingImpl(config)
+    if not isinstance(config, NvidiaPostTrainingConfig):
+        raise RuntimeError(f"Unexpected config type: {type(config)}")
+
+    impl = NvidiaPostTrainingAdapter(config)
     return impl
+
+
+__all__ = ["get_adapter_impl", "NvidiaPostTrainingAdapter"]

llama_stack/providers/remote/post_training/nvidia/post_training.py

@@ -50,7 +50,7 @@ class ListNvidiaPostTrainingJobs(BaseModel):
     data: List[NvidiaPostTrainingJob]
 
 
-class NvidiaPostTrainingImpl:
+class NvidiaPostTrainingAdapter:
     def __init__(self, config: NvidiaPostTrainingConfig):
         self.config = config
         self.headers = {}
@@ -226,12 +226,8 @@ class NvidiaPostTrainingImpl:
                 # Extract LoRA-specific parameters
                 lora_config = {k: v for k, v in algorithm_config.items() if k != "type"}
                 job_config["hyperparameters"]["lora"] = lora_config
-
-                # Add adapter_dim if available in training_config
-                if training_config.get("algorithm_config", {}).get("adapter_dim"):
-                    job_config["hyperparameters"]["lora"]["adapter_dim"] = training_config["algorithm_config"][
-                        "adapter_dim"
-                    ]
+            else:
+                raise NotImplementedError(f"Unsupported algorithm config: {algorithm_config}")
 
         # Create the customization job
         response = await self._make_request(