HunyuanImage21

src/diffusers/__init__.py

@@ -149,7 +149,9 @@
     _import_structure["guiders"].extend(
         [
             "AdaptiveProjectedGuidance",
+            "AdaptiveProjectedMixGuidance",
             "AutoGuidance",
+            "BaseGuidance",
             "ClassifierFreeGuidance",
             "ClassifierFreeZeroStarGuidance",
             "FrequencyDecoupledGuidance",
@@ -184,6 +186,8 @@
             "AutoencoderKLAllegro",
             "AutoencoderKLCogVideoX",
             "AutoencoderKLCosmos",
+            "AutoencoderKLHunyuanImage",
+            "AutoencoderKLHunyuanImageRefiner",
             "AutoencoderKLHunyuanVideo",
             "AutoencoderKLLTXVideo",
             "AutoencoderKLMagvit",
@@ -216,6 +220,7 @@
             "HunyuanDiT2DControlNetModel",
             "HunyuanDiT2DModel",
             "HunyuanDiT2DMultiControlNetModel",
+            "HunyuanImageTransformer2DModel",
             "HunyuanVideoFramepackTransformer3DModel",
             "HunyuanVideoTransformer3DModel",
             "I2VGenXLUNet",
@@ -460,6 +465,8 @@
             "HunyuanDiTControlNetPipeline",
             "HunyuanDiTPAGPipeline",
             "HunyuanDiTPipeline",
+            "HunyuanImagePipeline",
+            "HunyuanImageRefinerPipeline",
             "HunyuanSkyreelsImageToVideoPipeline",
             "HunyuanVideoFramepackPipeline",
             "HunyuanVideoImageToVideoPipeline",
@@ -845,7 +852,9 @@
     else:
         from .guiders import (
             AdaptiveProjectedGuidance,
+            AdaptiveProjectedMixGuidance,
             AutoGuidance,
+            BaseGuidance,
             ClassifierFreeGuidance,
             ClassifierFreeZeroStarGuidance,
             FrequencyDecoupledGuidance,
@@ -876,6 +885,8 @@
             AutoencoderKLAllegro,
             AutoencoderKLCogVideoX,
             AutoencoderKLCosmos,
+            AutoencoderKLHunyuanImage,
+            AutoencoderKLHunyuanImageRefiner,
             AutoencoderKLHunyuanVideo,
             AutoencoderKLLTXVideo,
             AutoencoderKLMagvit,
@@ -908,6 +919,7 @@
             HunyuanDiT2DControlNetModel,
             HunyuanDiT2DModel,
             HunyuanDiT2DMultiControlNetModel,
+            HunyuanImageTransformer2DModel,
             HunyuanVideoFramepackTransformer3DModel,
             HunyuanVideoTransformer3DModel,
             I2VGenXLUNet,
@@ -1122,6 +1134,8 @@
             HunyuanDiTControlNetPipeline,
             HunyuanDiTPAGPipeline,
             HunyuanDiTPipeline,
+            HunyuanImagePipeline,
+            HunyuanImageRefinerPipeline,
             HunyuanSkyreelsImageToVideoPipeline,
             HunyuanVideoFramepackPipeline,
             HunyuanVideoImageToVideoPipeline,

src/diffusers/guiders/__init__.py

@@ -19,23 +19,13 @@
 
 if is_torch_available():
     from .adaptive_projected_guidance import AdaptiveProjectedGuidance
+    from .adaptive_projected_guidance_mix import AdaptiveProjectedMixGuidance
     from .auto_guidance import AutoGuidance
     from .classifier_free_guidance import ClassifierFreeGuidance
     from .classifier_free_zero_star_guidance import ClassifierFreeZeroStarGuidance
     from .frequency_decoupled_guidance import FrequencyDecoupledGuidance
+    from .guider_utils import BaseGuidance
     from .perturbed_attention_guidance import PerturbedAttentionGuidance
     from .skip_layer_guidance import SkipLayerGuidance
     from .smoothed_energy_guidance import SmoothedEnergyGuidance
     from .tangential_classifier_free_guidance import TangentialClassifierFreeGuidance
-
-    GuiderType = Union[
-        AdaptiveProjectedGuidance,
-        AutoGuidance,
-        ClassifierFreeGuidance,
-        ClassifierFreeZeroStarGuidance,
-        FrequencyDecoupledGuidance,
-        PerturbedAttentionGuidance,
-        SkipLayerGuidance,
-        SmoothedEnergyGuidance,
-        TangentialClassifierFreeGuidance,
-    ]

src/diffusers/guiders/adaptive_projected_guidance.py

@@ -65,8 +65,9 @@ def __init__(
         use_original_formulation: bool = False,
         start: float = 0.0,
         stop: float = 1.0,
+        enabled: bool = True,
     ):
-        super().__init__(start, stop)
+        super().__init__(start, stop, enabled)
 
         self.guidance_scale = guidance_scale
         self.adaptive_projected_guidance_momentum = adaptive_projected_guidance_momentum
@@ -85,10 +86,10 @@ def prepare_inputs(
         if self._step == 0:
             if self.adaptive_projected_guidance_momentum is not None:
                 self.momentum_buffer = MomentumBuffer(self.adaptive_projected_guidance_momentum)
-        tuple_indices = [0] if self.num_conditions == 1 else [0, 1]
+        tuple_indices = [0] if self.num_conditions == 1 or not self._is_apg_enabled() else [0, 1]
         data_batches = []
-        for i in range(self.num_conditions):
-            data_batch = self._prepare_batch(input_fields, data, tuple_indices[i], self._input_predictions[i])
+        for tuple_idx, input_prediction in zip(tuple_indices, self._input_predictions):
+            data_batch = self._prepare_batch(input_fields, data, tuple_idx, input_prediction)
             data_batches.append(data_batch)
         return data_batches