Merge branch 'huggingface:main' into main

Author: oneflyingfish

src/diffusers/__init__.py

@@ -495,6 +495,7 @@
             "LTXImageToVideoPipeline",
             "LTXLatentUpsamplePipeline",
             "LTXPipeline",
+            "LucyEditPipeline",
             "Lumina2Pipeline",
             "Lumina2Text2ImgPipeline",
             "LuminaPipeline",
@@ -1149,6 +1150,7 @@
             LTXImageToVideoPipeline,
             LTXLatentUpsamplePipeline,
             LTXPipeline,
+            LucyEditPipeline,
             Lumina2Pipeline,
             Lumina2Text2ImgPipeline,
             LuminaPipeline,

src/diffusers/models/autoencoders/autoencoder_kl_wan.py

@@ -1052,7 +1052,7 @@ def __init__(
             is_residual=is_residual,
         )
 
-        self.spatial_compression_ratio = 2 ** len(self.temperal_downsample)
+        self.spatial_compression_ratio = scale_factor_spatial
 
         # When decoding a batch of video latents at a time, one can save memory by slicing across the batch dimension
         # to perform decoding of a single video latent at a time.
@@ -1145,12 +1145,13 @@ def clear_cache(self):
     def _encode(self, x: torch.Tensor):
         _, _, num_frame, height, width = x.shape
 
-        if self.use_tiling and (width > self.tile_sample_min_width or height > self.tile_sample_min_height):
-            return self.tiled_encode(x)
-
         self.clear_cache()
         if self.config.patch_size is not None:
             x = patchify(x, patch_size=self.config.patch_size)
+
+        if self.use_tiling and (width > self.tile_sample_min_width or height > self.tile_sample_min_height):
+            return self.tiled_encode(x)
+
         iter_ = 1 + (num_frame - 1) // 4
         for i in range(iter_):
             self._enc_conv_idx = [0]

src/diffusers/pipelines/__init__.py

@@ -285,6 +285,7 @@
     ]
     _import_structure["lumina"] = ["LuminaPipeline", "LuminaText2ImgPipeline"]
     _import_structure["lumina2"] = ["Lumina2Pipeline", "Lumina2Text2ImgPipeline"]
+    _import_structure["lucy"] = ["LucyEditPipeline"]
     _import_structure["marigold"].extend(
         [
             "MarigoldDepthPipeline",
@@ -682,6 +683,7 @@
             LEditsPPPipelineStableDiffusionXL,
         )
         from .ltx import LTXConditionPipeline, LTXImageToVideoPipeline, LTXLatentUpsamplePipeline, LTXPipeline
+        from .lucy import LucyEditPipeline
         from .lumina import LuminaPipeline, LuminaText2ImgPipeline
         from .lumina2 import Lumina2Pipeline, Lumina2Text2ImgPipeline
         from .marigold import (

src/diffusers/pipelines/lucy/__init__.py

@@ -0,0 +1,47 @@
+from typing import TYPE_CHECKING
+
+from ...utils import (
+    DIFFUSERS_SLOW_IMPORT,
+    OptionalDependencyNotAvailable,
+    _LazyModule,
+    get_objects_from_module,
+    is_torch_available,
+    is_transformers_available,
+)
+
+
+_dummy_objects = {}
+_import_structure = {}
+
+
+try:
+    if not (is_transformers_available() and is_torch_available()):
+        raise OptionalDependencyNotAvailable()
+except OptionalDependencyNotAvailable:
+    from ...utils import dummy_torch_and_transformers_objects  # noqa F403
+
+    _dummy_objects.update(get_objects_from_module(dummy_torch_and_transformers_objects))
+else:
+    _import_structure["pipeline_lucy_edit"] = ["LucyEditPipeline"]
+if TYPE_CHECKING or DIFFUSERS_SLOW_IMPORT:
+    try:
+        if not (is_transformers_available() and is_torch_available()):
+            raise OptionalDependencyNotAvailable()
+
+    except OptionalDependencyNotAvailable:
+        from ...utils.dummy_torch_and_transformers_objects import *
+    else:
+        from .pipeline_lucy_edit import LucyEditPipeline
+
+else:
+    import sys
+
+    sys.modules[__name__] = _LazyModule(
+        __name__,
+        globals()["__file__"],
+        _import_structure,
+        module_spec=__spec__,
+    )
+
+    for name, value in _dummy_objects.items():
+        setattr(sys.modules[__name__], name, value)