feat: make training config fields optional (#1861)

# What does this PR do?

Today, supervised_fine_tune itself and the `TrainingConfig` class have a
bunch of required fields that a provider implementation might not need.

for example, if a provider wants to handle hyperparameters in its
configuration as well as any type of dataset retrieval, optimizer or
LoRA config, a user will still need to pass in a virtually empty
`DataConfig`, `OptimizerConfig` and `AlgorithmConfig` in some cases.

Many of these fields are intended to work specifically with llama models
and knobs intended for customizing inline.

Adding remote post_training providers will require loosening these
arguments, or forcing users to pass in empty objects to satisfy the
pydantic models.

Signed-off-by: Charlie Doern <cdoern@redhat.com>

llama_stack/providers/inline/post_training/torchtune/recipes/lora_finetuning_single_device.py

@@ -38,6 +38,8 @@ from llama_stack.apis.datasetio import DatasetIO
 from llama_stack.apis.datasets import Datasets
 from llama_stack.apis.post_training import (
     Checkpoint,
+    DataConfig,
+    EfficiencyConfig,
     LoraFinetuningConfig,
     OptimizerConfig,
     QATFinetuningConfig,
@@ -89,6 +91,10 @@ class LoraFinetuningSingleDevice:
         datasetio_api: DatasetIO,
         datasets_api: Datasets,
     ) -> None:
+        assert isinstance(training_config.data_config, DataConfig), "DataConfig must be initialized"
+
+        assert isinstance(training_config.efficiency_config, EfficiencyConfig), "EfficiencyConfig must be initialized"
+
         self.job_uuid = job_uuid
         self.training_config = training_config
         if not isinstance(algorithm_config, LoraFinetuningConfig):
@@ -188,6 +194,7 @@ class LoraFinetuningSingleDevice:
         self._tokenizer = await self._setup_tokenizer()
         log.info("Tokenizer is initialized.")
 
+        assert isinstance(self.training_config.optimizer_config, OptimizerConfig), "OptimizerConfig must be initialized"
         self._optimizer = await self._setup_optimizer(optimizer_config=self.training_config.optimizer_config)
         log.info("Optimizer is initialized.")
 
@@ -195,6 +202,8 @@ class LoraFinetuningSingleDevice:
         self._model.set_num_output_chunks(self._loss_fn.num_output_chunks)
         log.info("Loss is initialized.")
 
+        assert isinstance(self.training_config.data_config, DataConfig), "DataConfig must be initialized"
+
         self._training_sampler, self._training_dataloader = await self._setup_data(
             dataset_id=self.training_config.data_config.dataset_id,
             tokenizer=self._tokenizer,
@@ -452,6 +461,7 @@ class LoraFinetuningSingleDevice:
         """
         The core training loop.
         """
+        assert isinstance(self.training_config.data_config, DataConfig), "DataConfig must be initialized"
         # Initialize tokens count and running loss (for grad accumulation)
         t0 = time.perf_counter()
         running_loss: float = 0.0