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

tests/backend/model_manager_2/util/test_hf_model_select.py

@@ -192,6 +192,7 @@ def sdxl_base_files() -> List[Path]:
                 "text_encoder/model.onnx",
                 "text_encoder_2/config.json",
                 "text_encoder_2/model.onnx",
+                "text_encoder_2/model.onnx_data",
                 "tokenizer/merges.txt",
                 "tokenizer/special_tokens_map.json",
                 "tokenizer/tokenizer_config.json",
@@ -202,6 +203,7 @@ def sdxl_base_files() -> List[Path]:
                 "tokenizer_2/vocab.json",
                 "unet/config.json",
                 "unet/model.onnx",
+                "unet/model.onnx_data",
                 "vae_decoder/config.json",
                 "vae_decoder/model.onnx",
                 "vae_encoder/config.json",