remove deprecated LoRA blocks

my_pipeline.py

@@ -25,12 +25,7 @@ from diffusers.loaders import (
     TextualInversionLoaderMixin,
 )
 from diffusers.models import AutoencoderKL, UNet2DConditionModel
-from diffusers.models.attention_processor import (
-    AttnProcessor2_0,
-    LoRAAttnProcessor2_0,
-    LoRAXFormersAttnProcessor,
-    XFormersAttnProcessor,
-)
+from diffusers.models.attention_processor import AttnProcessor2_0, XFormersAttnProcessor
 from diffusers.models.lora import adjust_lora_scale_text_encoder
 from diffusers.schedulers import KarrasDiffusionSchedulers
 from diffusers.utils import (
@@ -135,6 +130,7 @@ class MyPipeline(
             watermark output images. If not defined, it will default to True if the package is installed, otherwise no
             watermarker will be used.
     """
+
     model_cpu_offload_seq = "text_encoder->text_encoder_2->unet->vae"
     _optional_components = ["tokenizer", "tokenizer_2", "text_encoder", "text_encoder_2"]
 
@@ -571,12 +567,7 @@ class MyPipeline(
         self.vae.to(dtype=torch.float32)
         use_torch_2_0_or_xformers = isinstance(
             self.vae.decoder.mid_block.attentions[0].processor,
-            (
-                AttnProcessor2_0,
-                XFormersAttnProcessor,
-                LoRAXFormersAttnProcessor,
-                LoRAAttnProcessor2_0,
-            ),
+            (AttnProcessor2_0, XFormersAttnProcessor),
         )
         # if xformers or torch_2_0 is used attention block does not need
         # to be in float32 which can save lots of memory
@@ -971,4 +962,4 @@ class MyPipeline(
         # Offload all models
         self.maybe_free_model_hooks()
 
-        return (image,)
+        return (image,)

unet/my_unet_model.py

@@ -34,7 +34,7 @@ from diffusers.models.embeddings import (
     ImageHintTimeEmbedding,
     ImageProjection,
     ImageTimeEmbedding,
-    PositionNet,
+    GLIGENTextBoundingboxProjection,
     TextImageProjection,
     TextImageTimeEmbedding,
     TextTimeEmbedding,
@@ -42,7 +42,7 @@ from diffusers.models.embeddings import (
     Timesteps,
 )
 from diffusers.models.modeling_utils import ModelMixin
-from diffusers.models.unet_2d_blocks import (
+from diffusers.models.unets.unet_2d_blocks import (
     UNetMidBlock2DCrossAttn,
     UNetMidBlock2DSimpleCrossAttn,
     get_down_block,
@@ -586,7 +586,7 @@ class MyUNetModel(ModelMixin, ConfigMixin, UNet2DConditionLoadersMixin):
                 positive_len = cross_attention_dim[0]
 
             feature_type = "text-only" if attention_type == "gated" else "text-image"
-            self.position_net = PositionNet(
+            self.position_net = GLIGENTextBoundingboxProjection(
                 positive_len=positive_len, out_dim=cross_attention_dim, feature_type=feature_type
             )