stylestudio

ip_adapter/__init__.py

@@ -1,15 +1,12 @@
-from .ip_adapter import IPAdapter, IPAdapterPlus, IPAdapterPlusXL, IPAdapterXL, IPAdapterFull,IPAdapterXL_CS,IPAdapter_CS
+from .ip_adapter import IPAdapter, IPAdapterPlus, IPAdapterPlusXL, IPAdapterXL, IPAdapterFull
 from .ip_adapter import CSGO
-from .ip_adapter import StyleStudio_Adapter, StyleStudio_Adapter_exp
-from .ip_adapter import IPAdapterXL_cross_modal
+from .ip_adapter import StyleStudio_Adapter
 __all__ = [
     "IPAdapter",
     "IPAdapterPlus",
     "IPAdapterPlusXL",
     "IPAdapterXL",
+    "IPAdapterFull",
     "CSGO",
     "StyleStudio_Adapter",
-    "StyleStudio_Adapter_exp",
-    "IPAdapterXL_cross_modal",
-    "IPAdapterFull",
 ]

ip_adapter/ip_adapter.py

@@ -246,6 +246,35 @@ class IPAdapter:
 
         return images
 
+class IPAdapterPlus(IPAdapter):
+    """IP-Adapter with fine-grained features"""
+
+    def init_proj(self):
+        image_proj_model = Resampler(
+            dim=self.pipe.unet.config.cross_attention_dim,
+            depth=4,
+            dim_head=64,
+            heads=12,
+            num_queries=self.num_tokens,
+            embedding_dim=self.image_encoder.config.hidden_size,
+            output_dim=self.pipe.unet.config.cross_attention_dim,
+            ff_mult=4,
+        ).to(self.device, dtype=torch.float16)
+        return image_proj_model
+
+    @torch.inference_mode()
+    def get_image_embeds(self, pil_image=None, clip_image_embeds=None):
+        if isinstance(pil_image, Image.Image):
+            pil_image = [pil_image]
+        clip_image = self.clip_image_processor(images=pil_image, return_tensors="pt").pixel_values
+        clip_image = clip_image.to(self.device, dtype=torch.float16)
+        clip_image_embeds = self.image_encoder(clip_image, output_hidden_states=True).hidden_states[-2]
+        image_prompt_embeds = self.image_proj_model(clip_image_embeds)
+        uncond_clip_image_embeds = self.image_encoder(
+            torch.zeros_like(clip_image), output_hidden_states=True
+        ).hidden_states[-2]
+        uncond_image_prompt_embeds = self.image_proj_model(uncond_clip_image_embeds)
+        return image_prompt_embeds, uncond_image_prompt_embeds
 
 class IPAdapter_CS:
     def __init__(self, sd_pipe, image_encoder_path, ip_ckpt, device, num_content_tokens=4,