BREAKING CHANGES: invocations now require model key, not base/type/name

- Implement new model loader and modify invocations and embeddings

- Finish implementation loaders for all models currently supported by
  InvokeAI.

- Move lora, textual_inversion, and model patching support into
  backend/embeddings.

- Restore support for model cache statistics collection (a little ugly,
  needs work).

- Fixed up invocations that load and patch models.

- Move seamless and silencewarnings utils into better location

invokeai/backend/model_manager/config.py

@@ -30,8 +30,11 @@ from typing_extensions import Annotated, Any, Dict
 
 from invokeai.backend.onnx.onnx_runtime import IAIOnnxRuntimeModel
 
+from ..embeddings.embedding_base import EmbeddingModelRaw
 from ..ip_adapter.ip_adapter import IPAdapter, IPAdapterPlus
 
+AnyModel = Union[ModelMixin, torch.nn.Module, IAIOnnxRuntimeModel, IPAdapter, IPAdapterPlus, EmbeddingModelRaw]
+
 
 class InvalidModelConfigException(Exception):
     """Exception for when config parser doesn't recognized this combination of model type and format."""
@@ -299,7 +302,7 @@ AnyModelConfig = Union[
 ]
 
 AnyModelConfigValidator = TypeAdapter(AnyModelConfig)
-AnyModel = Union[ModelMixin, torch.nn.Module, IAIOnnxRuntimeModel, IPAdapter, IPAdapterPlus]
+
 
 # IMPLEMENTATION NOTE:
 # The preferred alternative to the above is a discriminated Union as shown