Add FLUX IP-Adapter starter models.

invokeai/app/invocations/flux_ip_adapter.py

@@ -68,8 +68,9 @@ class FluxIPAdapterInvocation(BaseInvocation):
         assert isinstance(ip_adapter_info, (IPAdapterInvokeAIConfig, IPAdapterCheckpointConfig))
 
         # Note: There is a IPAdapterInvokeAIConfig.image_encoder_model_id field, but it isn't trustworthy.
-        image_encoder_model_id, image_encoder_model_name = CLIP_VISION_MODEL_MAP[self.clip_vision_model]
-
+        image_encoder_starter_model = CLIP_VISION_MODEL_MAP[self.clip_vision_model]
+        image_encoder_model_id = image_encoder_starter_model.source
+        image_encoder_model_name = image_encoder_starter_model.name
         image_encoder_model = IPAdapterInvocation.get_clip_image_encoder(
             context, image_encoder_model_id, image_encoder_model_name
         )

invokeai/app/invocations/ip_adapter.py

@@ -18,6 +18,12 @@ from invokeai.backend.model_manager.config import (
     IPAdapterInvokeAIConfig,
     ModelType,
 )
+from invokeai.backend.model_manager.starter_models import (
+    StarterModel,
+    clip_vit_l_image_encoder,
+    ip_adapter_sd_image_encoder,
+    ip_adapter_sdxl_image_encoder,
+)
 
 
 class IPAdapterField(BaseModel):
@@ -56,10 +62,10 @@ class IPAdapterOutput(BaseInvocationOutput):
     ip_adapter: IPAdapterField = OutputField(description=FieldDescriptions.ip_adapter, title="IP-Adapter")
 
 
-CLIP_VISION_MODEL_MAP = {
-    "ViT-L": ("InvokeAI/clip-vit-large-patch14", "clip-vit-large-patch14"),
-    "ViT-H": ("InvokeAI/ip_adapter_sd_image_encoder", "ip_adapter_sd_image_encoder"),
-    "ViT-G": ("InvokeAI/ip_adapter_sdxl_image_encoder", "ip_adapter_sdxl_image_encoder"),
+CLIP_VISION_MODEL_MAP: dict[Literal["ViT-L", "ViT-H", "ViT-G"], StarterModel] = {
+    "ViT-L": clip_vit_l_image_encoder,
+    "ViT-H": ip_adapter_sd_image_encoder,
+    "ViT-G": ip_adapter_sdxl_image_encoder,
 }
 
 
@@ -116,7 +122,9 @@ class IPAdapterInvocation(BaseInvocation):
             image_encoder_model_id = ip_adapter_info.image_encoder_model_id
             image_encoder_model_name = image_encoder_model_id.split("/")[-1].strip()
         else:
-            image_encoder_model_id, image_encoder_model_name = CLIP_VISION_MODEL_MAP[self.clip_vision_model]
+            image_encoder_starter_model = CLIP_VISION_MODEL_MAP[self.clip_vision_model]
+            image_encoder_model_id = image_encoder_starter_model.source
+            image_encoder_model_name = image_encoder_starter_model.name
 
         image_encoder_model = self.get_clip_image_encoder(context, image_encoder_model_id, image_encoder_model_name)
 

invokeai/backend/model_manager/starter_models.py

@@ -25,22 +25,6 @@ class StarterModelBundles(BaseModel):
     models: list[StarterModel]
 
 
-ip_adapter_sd_image_encoder = StarterModel(
-    name="IP Adapter SD1.5 Image Encoder",
-    base=BaseModelType.StableDiffusion1,
-    source="InvokeAI/ip_adapter_sd_image_encoder",
-    description="IP Adapter SD Image Encoder",
-    type=ModelType.CLIPVision,
-)
-
-ip_adapter_sdxl_image_encoder = StarterModel(
-    name="IP Adapter SDXL Image Encoder",
-    base=BaseModelType.StableDiffusionXL,
-    source="InvokeAI/ip_adapter_sdxl_image_encoder",
-    description="IP Adapter SDXL Image Encoder",
-    type=ModelType.CLIPVision,
-)
-
 cyberrealistic_negative = StarterModel(
     name="CyberRealistic Negative v3",
     base=BaseModelType.StableDiffusion1,
@@ -49,6 +33,32 @@ cyberrealistic_negative = StarterModel(
     type=ModelType.TextualInversion,
 )
 
+# region CLIP Image Encoders
+ip_adapter_sd_image_encoder = StarterModel(
+    name="IP Adapter SD1.5 Image Encoder",
+    base=BaseModelType.StableDiffusion1,
+    source="InvokeAI/ip_adapter_sd_image_encoder",
+    description="IP Adapter SD Image Encoder",
+    type=ModelType.CLIPVision,
+)
+ip_adapter_sdxl_image_encoder = StarterModel(
+    name="IP Adapter SDXL Image Encoder",
+    base=BaseModelType.StableDiffusionXL,
+    source="InvokeAI/ip_adapter_sdxl_image_encoder",
+    description="IP Adapter SDXL Image Encoder",
+    type=ModelType.CLIPVision,
+)
+# Note: This model is installed from the same source as the CLIPEmbed model below. The model contains both the image
+# encoder and the text encoder, but we need separate model entries so that they get loaded correctly.
+clip_vit_l_image_encoder = StarterModel(
+    name="clip-vit-large-patch14",
+    base=BaseModelType.Any,
+    source="InvokeAI/clip-vit-large-patch14",
+    description="CLIP ViT-L Image Encoder",
+    type=ModelType.CLIPVision,
+)
+# endregion
+
 # region TextEncoders
 t5_base_encoder = StarterModel(
     name="t5_base_encoder",
@@ -254,6 +264,14 @@ ip_adapter_sdxl = StarterModel(
     type=ModelType.IPAdapter,
     dependencies=[ip_adapter_sdxl_image_encoder],
 )
+ip_adapter_flux = StarterModel(
+    name="XLabs FLUX IP-Adapter",
+    base=BaseModelType.Flux,
+    source="https://huggingface.co/XLabs-AI/flux-ip-adapter/resolve/main/flux-ip-adapter.safetensors",
+    description="FLUX IP-Adapter",
+    type=ModelType.IPAdapter,
+    dependencies=[clip_vit_l_image_encoder],
+)
 # endregion
 # region ControlNet
 qr_code_cnet_sd1 = StarterModel(
@@ -555,6 +573,7 @@ STARTER_MODELS: list[StarterModel] = [
     ip_adapter_plus_sd1,
     ip_adapter_plus_face_sd1,
     ip_adapter_sdxl,
+    ip_adapter_flux,
     qr_code_cnet_sd1,
     qr_code_cnet_sdxl,
     canny_sd1,
@@ -642,6 +661,7 @@ flux_bundle: list[StarterModel] = [
     t5_8b_quantized_encoder,
     clip_l_encoder,
     union_cnet_flux,
+    ip_adapter_flux,
 ]
 
 STARTER_BUNDLES: dict[str, list[StarterModel]] = {