Merge branch 'main' into qwenimage-edit-plus-followup

Author: sayakpaul

tests/pipelines/kandinsky/test_kandinsky.py

@@ -18,11 +18,13 @@
 import unittest
 
 import numpy as np
+import pytest
 import torch
 from transformers import XLMRobertaTokenizerFast
 
 from diffusers import DDIMScheduler, KandinskyPipeline, KandinskyPriorPipeline, UNet2DConditionModel, VQModel
 from diffusers.pipelines.kandinsky.text_encoder import MCLIPConfig, MultilingualCLIP
+from diffusers.utils import is_transformers_version
 
 from ...testing_utils import (
     backend_empty_cache,
@@ -215,6 +217,9 @@ def get_dummy_inputs(self, device, seed=0):
         dummy = Dummies()
         return dummy.get_dummy_inputs(device=device, seed=seed)
 
+    @pytest.mark.xfail(
+        condition=is_transformers_version(">=", "4.56.2"), reason="Latest transformers changes the slices", strict=True
+    )
     def test_kandinsky(self):
         device = "cpu"
 

tests/pipelines/kandinsky/test_kandinsky_combined.py

@@ -16,8 +16,10 @@
 import unittest
 
 import numpy as np
+import pytest
 
 from diffusers import KandinskyCombinedPipeline, KandinskyImg2ImgCombinedPipeline, KandinskyInpaintCombinedPipeline
+from diffusers.utils import is_transformers_version
 
 from ...testing_utils import enable_full_determinism, require_torch_accelerator, torch_device
 from ..test_pipelines_common import PipelineTesterMixin
@@ -73,6 +75,9 @@ def get_dummy_inputs(self, device, seed=0):
         )
         return inputs
 
+    @pytest.mark.xfail(
+        condition=is_transformers_version(">=", "4.56.2"), reason="Latest transformers changes the slices", strict=True
+    )
     def test_kandinsky(self):
         device = "cpu"
 
@@ -181,6 +186,9 @@ def get_dummy_inputs(self, device, seed=0):
         inputs.pop("negative_image_embeds")
         return inputs
 
+    @pytest.mark.xfail(
+        condition=is_transformers_version(">=", "4.56.2"), reason="Latest transformers changes the slices", strict=True
+    )
     def test_kandinsky(self):
         device = "cpu"
 
@@ -292,6 +300,9 @@ def get_dummy_inputs(self, device, seed=0):
         inputs.pop("negative_image_embeds")
         return inputs
 
+    @pytest.mark.xfail(
+        condition=is_transformers_version(">=", "4.56.2"), reason="Latest transformers changes the slices", strict=True
+    )
     def test_kandinsky(self):
         device = "cpu"
 

tests/pipelines/kandinsky/test_kandinsky_img2img.py

@@ -18,6 +18,7 @@
 import unittest
 
 import numpy as np
+import pytest
 import torch
 from PIL import Image
 from transformers import XLMRobertaTokenizerFast
@@ -31,6 +32,7 @@
     VQModel,
 )
 from diffusers.pipelines.kandinsky.text_encoder import MCLIPConfig, MultilingualCLIP
+from diffusers.utils import is_transformers_version
 
 from ...testing_utils import (
     backend_empty_cache,
@@ -237,6 +239,9 @@ def get_dummy_inputs(self, device, seed=0):
         dummies = Dummies()
         return dummies.get_dummy_inputs(device=device, seed=seed)
 
+    @pytest.mark.xfail(
+        condition=is_transformers_version(">=", "4.56.2"), reason="Latest transformers changes the slices", strict=True
+    )
     def test_kandinsky_img2img(self):
         device = "cpu"
 

tests/pipelines/kandinsky/test_kandinsky_inpaint.py

@@ -18,12 +18,14 @@
 import unittest
 
 import numpy as np
+import pytest
 import torch
 from PIL import Image
 from transformers import XLMRobertaTokenizerFast
 
 from diffusers import DDIMScheduler, KandinskyInpaintPipeline, KandinskyPriorPipeline, UNet2DConditionModel, VQModel
 from diffusers.pipelines.kandinsky.text_encoder import MCLIPConfig, MultilingualCLIP
+from diffusers.utils import is_transformers_version
 
 from ...testing_utils import (
     backend_empty_cache,
@@ -231,6 +233,9 @@ def get_dummy_inputs(self, device, seed=0):
         dummies = Dummies()
         return dummies.get_dummy_inputs(device=device, seed=seed)
 
+    @pytest.mark.xfail(
+        condition=is_transformers_version(">=", "4.56.2"), reason="Latest transformers changes the slices", strict=True
+    )
     def test_kandinsky_inpaint(self):
         device = "cpu"