Rename DiffusersControlNetFlux -> InstantXControlNetFlux.

invokeai/backend/flux/controlnet/instantx_controlnet_flux.py

@@ -18,7 +18,7 @@ from invokeai.backend.flux.modules.layers import (
 
 
 @dataclass
-class DiffusersControlNetFluxOutput:
+class InstantXControlNetFluxOutput:
     controlnet_block_samples: list[torch.Tensor] | None
     controlnet_single_block_samples: list[torch.Tensor] | None
 
@@ -36,7 +36,7 @@ class DiffusersControlNetFluxOutput:
 # - axes_dims_rope: axes_dim
 
 
-class DiffusersControlNetFlux(torch.nn.Module):
+class InstantXControlNetFlux(torch.nn.Module):
     def __init__(self, params: FluxParams, num_control_modes: int | None = None):
         """
         Args:
@@ -114,7 +114,7 @@ class DiffusersControlNetFlux(torch.nn.Module):
         timesteps: torch.Tensor,
         y: torch.Tensor,
         guidance: torch.Tensor | None = None,
-    ) -> DiffusersControlNetFluxOutput:
+    ) -> InstantXControlNetFluxOutput:
         if img.ndim != 3 or txt.ndim != 3:
             raise ValueError("Input img and txt tensors must have 3 dimensions.")
 
@@ -168,7 +168,7 @@ class DiffusersControlNetFlux(torch.nn.Module):
             single_block_sample = controlnet_block(single_block_sample)
             controlnet_single_block_samples.append(single_block_sample)
 
-        return DiffusersControlNetFluxOutput(
+        return InstantXControlNetFluxOutput(
             controlnet_block_samples=controlnet_double_block_samples or None,
             controlnet_single_block_samples=controlnet_single_block_samples or None,
         )

invokeai/backend/flux/controlnet/state_dict_utils.py

@@ -173,11 +173,11 @@ def _convert_flux_single_block_sd_from_diffusers_to_bfl_format(
 
 def convert_diffusers_instantx_state_dict_to_bfl_format(sd: Dict[str, torch.Tensor]) -> Dict[str, torch.Tensor]:
     """Convert an InstantX ControlNet state dict to the format that can be loaded by our internal
-    DiffusersControlNetFlux.
+    InstantXControlNetFlux model.
 
     The original InstantX ControlNet model was developed to be used in diffusers. We have ported the original
-    implementation to DiffusersControlNetFlux to make it compatible with BFL-style models. This function converts the
-    original state dict to the format expected by DiffusersControlNetFlux.
+    implementation to InstantXControlNetFlux to make it compatible with BFL-style models. This function converts the
+    original state dict to the format expected by InstantXControlNetFlux.
     """
     # Shallow copy sd so that we can pop keys from it without modifying the original.
     sd = sd.copy()

invokeai/backend/model_manager/load/model_loaders/flux.py

@@ -10,7 +10,7 @@ from safetensors.torch import load_file
 from transformers import AutoConfig, AutoModelForTextEncoding, CLIPTextModel, CLIPTokenizer, T5EncoderModel, T5Tokenizer
 
 from invokeai.app.services.config.config_default import get_config
-from invokeai.backend.flux.controlnet.diffusers_controlnet_flux import DiffusersControlNetFlux
+from invokeai.backend.flux.controlnet.instantx_controlnet_flux import InstantXControlNetFlux
 from invokeai.backend.flux.controlnet.state_dict_utils import (
     convert_diffusers_instantx_state_dict_to_bfl_format,
     infer_flux_params_from_state_dict,
@@ -341,7 +341,7 @@ class FluxControlnetModel(ModelLoader):
         num_control_modes = infer_instantx_num_control_modes_from_state_dict(sd)
 
         with accelerate.init_empty_weights():
-            model = DiffusersControlNetFlux(flux_params, num_control_modes)
+            model = InstantXControlNetFlux(flux_params, num_control_modes)
 
         model.load_state_dict(sd, assign=True)
         return model

tests/backend/flux/controlnet/test_state_dict_utils.py

@@ -1,7 +1,7 @@
 import pytest
 import torch
 
-from invokeai.backend.flux.controlnet.diffusers_controlnet_flux import DiffusersControlNetFlux
+from invokeai.backend.flux.controlnet.instantx_controlnet_flux import InstantXControlNetFlux
 from invokeai.backend.flux.controlnet.state_dict_utils import (
     convert_diffusers_instantx_state_dict_to_bfl_format,
     infer_flux_params_from_state_dict,
@@ -89,7 +89,7 @@ def test_load_instantx_from_state_dict():
     num_control_modes = infer_instantx_num_control_modes_from_state_dict(sd)
 
     with torch.device("meta"):
-        model = DiffusersControlNetFlux(flux_params, num_control_modes)
+        model = InstantXControlNetFlux(flux_params, num_control_modes)
 
     model_sd = model.state_dict()