feat: Support bin IP Adapter files

invokeai/backend/ip_adapter/ip_adapter.py

@@ -221,6 +221,8 @@ def load_ip_adapter_tensors(ip_adapter_ckpt_path: pathlib.Path, device: str) ->
                 state_dict["ip_adapter"][key.replace("ip_adapter.", "")] = model[key]
             else:
                 raise RuntimeError(f"Encountered unexpected IP Adapter state dict key: '{key}'.")
+    elif ip_adapter_ckpt_path.suffix == ".bin":
+        state_dict = torch.load(ip_adapter_ckpt_path, map_location="cpu")
     else:
         ip_adapter_diffusers_checkpoint_path = ip_adapter_ckpt_path / "ip_adapter.bin"
         state_dict = torch.load(ip_adapter_diffusers_checkpoint_path, map_location="cpu")

invokeai/backend/model_manager/probe.py

@@ -231,7 +231,7 @@ def get_model_type_from_checkpoint(cls, model_path: Path, checkpoint: Optional[C
                 return ModelType.LoRA
             elif any(key.startswith(v) for v in {"controlnet", "control_model", "input_blocks"}):
                 return ModelType.ControlNet
-            elif any(key.startswith(v) for v in {"image_proj.", "ip_adapter."}):
+            elif any(key.startswith(v) for v in {"image_proj", "ip_adapter"}):
                 return ModelType.IPAdapter
             elif key in {"emb_params", "string_to_param"}:
                 return ModelType.TextualInversion
@@ -542,9 +542,14 @@ class IPAdapterCheckpointProbe(CheckpointProbeBase):
     def get_base_type(self) -> BaseModelType:
         checkpoint = self.checkpoint
         for key in checkpoint.keys():
-            if not key.startswith(("image_proj.", "ip_adapter.")):
+            if not key.startswith(("image_proj", "ip_adapter")):
                 continue
-            cross_attention_dim = checkpoint["ip_adapter.1.to_k_ip.weight"].shape[-1]
+
+            if key in ["image_proj", "ip_adapter"]:
+                cross_attention_dim = checkpoint["ip_adapter"]["1.to_k_ip.weight"].shape[-1]
+            else:
+                cross_attention_dim = checkpoint["ip_adapter.1.to_k_ip.weight"].shape[-1]
+
             if cross_attention_dim == 768:
                 return BaseModelType.StableDiffusion1
             elif cross_attention_dim == 1024: