update

.github/workflows/pr_tests.yml

@@ -92,9 +92,8 @@ jobs:
             runner: aws-general-8-plus
             image: diffusers/diffusers-pytorch-cpu
             report: torch_example_cpu
-        transformers_version: ["main"]
 
-    name: ${{ matrix.config.name }} (transformers ${{ matrix.transformers_version }})
+    name: ${{ matrix.config.name }}
 
     runs-on:
       group: ${{ matrix.config.runner }}
@@ -116,11 +115,8 @@ jobs:
     - name: Install dependencies
       run: |
         uv pip install -e ".[quality]"
-        if [ "${{ matrix.transformers_version }}" = "main" ]; then
-          uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
-        else
-          uv pip uninstall transformers huggingface_hub && uv pip install transformers==${{ matrix.transformers_version }}
-        fi
+        #uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
+        uv pip uninstall transformers huggingface_hub && uv pip install transformers==4.57.1
         uv pip uninstall accelerate && uv pip install -U accelerate@git+https://github.com/huggingface/accelerate.git --no-deps
 
     - name: Environment
@@ -159,7 +155,7 @@ jobs:
       if: ${{ always() }}
       uses: actions/upload-artifact@v6
       with:
-        name: pr_${{ matrix.config.framework }}_${{ matrix.config.report }}_transformers_${{ matrix.transformers_version }}_test_reports
+        name: pr_${{ matrix.config.framework }}_${{ matrix.config.report }}_test_reports
         path: reports
 
   run_staging_tests:
@@ -224,10 +220,8 @@ jobs:
     needs: [check_code_quality, check_repository_consistency]
     strategy:
       fail-fast: false
-      matrix:
-        transformers_version: ["main"]
 
-    name: LoRA tests with PEFT main (transformers ${{ matrix.transformers_version }})
+    name: LoRA tests with PEFT main
 
     runs-on:
       group: aws-general-8-plus
@@ -253,12 +247,9 @@ jobs:
         uv pip install -U peft@git+https://github.com/huggingface/peft.git --no-deps
         uv pip install -U tokenizers
         uv pip uninstall accelerate && uv pip install -U accelerate@git+https://github.com/huggingface/accelerate.git --no-deps
-        if [ "${{ matrix.transformers_version }}" = "main" ]; then
-          uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
-        else
-          uv pip uninstall transformers huggingface_hub && uv pip install transformers==${{ matrix.transformers_version }}
-        fi
-        
+        #uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
+        uv pip uninstall transformers huggingface_hub && uv pip install transformers==4.57.1
+
     - name: Environment
       run: |
         python utils/print_env.py
@@ -284,6 +275,6 @@ jobs:
       if: ${{ always() }}
       uses: actions/upload-artifact@v6
       with:
-        name: pr_lora_transformers_${{ matrix.transformers_version }}_test_reports
+        name: pr_main_test_reports
         path: reports
 

.github/workflows/pr_tests_gpu.yml

@@ -14,7 +14,6 @@ on:
       - "tests/pipelines/test_pipelines_common.py"
       - "tests/models/test_modeling_common.py"
       - "examples/**/*.py"
-      - ".github/**.yml"
   workflow_dispatch:
 
 concurrency:
@@ -107,14 +106,13 @@ jobs:
           path: reports
 
   torch_pipelines_cuda_tests:
-    name: Torch Pipelines CUDA Tests (transformers ${{ matrix.transformers_version }})
+    name: Torch Pipelines CUDA Tests
     needs: setup_torch_cuda_pipeline_matrix
     strategy:
       fail-fast: false
       max-parallel: 8
       matrix:
         module: ${{ fromJson(needs.setup_torch_cuda_pipeline_matrix.outputs.pipeline_test_matrix) }}
-        transformers_version: ["main"]
     runs-on:
       group: aws-g4dn-2xlarge
     container:
@@ -133,12 +131,8 @@ jobs:
         run: |
           uv pip install -e ".[quality]"
           uv pip uninstall accelerate && uv pip install -U accelerate@git+https://github.com/huggingface/accelerate.git
-          if [ "${{ matrix.transformers_version }}" = "main" ]; then
-            uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
-          else
-            uv pip uninstall transformers huggingface_hub && uv pip install transformers==${{ matrix.transformers_version }}
-          fi
-          
+          #uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
+          uv pip uninstall transformers huggingface_hub && uv pip install transformers==4.57.1
 
       - name: Environment
         run: |
@@ -178,11 +172,11 @@ jobs:
         if: ${{ always() }}
         uses: actions/upload-artifact@v6
         with:
-          name: pipeline_${{ matrix.module }}_transformers_${{ matrix.transformers_version }}_test_reports
+          name: pipeline_${{ matrix.module }}_test_reports
           path: reports
 
   torch_cuda_tests:
-    name: Torch CUDA Tests (transformers ${{ matrix.transformers_version }})
+    name: Torch CUDA Tests
     needs: [check_code_quality, check_repository_consistency]
     runs-on:
       group: aws-g4dn-2xlarge
@@ -197,7 +191,6 @@ jobs:
       max-parallel: 4
       matrix:
         module: [models, schedulers, lora, others]
-        transformers_version: ["main"]
     steps:
     - name: Checkout diffusers
       uses: actions/checkout@v6
@@ -209,12 +202,8 @@ jobs:
         uv pip install -e ".[quality]"
         uv pip install peft@git+https://github.com/huggingface/peft.git
         uv pip uninstall accelerate && uv pip install -U accelerate@git+https://github.com/huggingface/accelerate.git
-        if [ "${{ matrix.transformers_version }}" = "main" ]; then
-          uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
-        else
-          uv pip uninstall transformers huggingface_hub && uv pip install transformers==${{ matrix.transformers_version }}
-        fi
-        
+        #uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
+        uv pip uninstall transformers huggingface_hub && uv pip install transformers==4.57.1
 
     - name: Environment
       run: |
@@ -252,16 +241,12 @@ jobs:
       if: ${{ always() }}
       uses: actions/upload-artifact@v6
       with:
-        name: torch_cuda_test_reports_${{ matrix.module }}_transformers_${{ matrix.transformers_version }}
+        name: torch_cuda_test_reports_${{ matrix.module }}
         path: reports
 
   run_examples_tests:
-    name: Examples PyTorch CUDA tests on Ubuntu (transformers ${{ matrix.transformers_version }})
+    name: Examples PyTorch CUDA tests on Ubuntu
     needs: [check_code_quality, check_repository_consistency]
-    strategy:
-      fail-fast: false
-      matrix:
-        transformers_version: ["main"]
     runs-on:
       group: aws-g4dn-2xlarge
 
@@ -279,11 +264,8 @@ jobs:
         nvidia-smi
     - name: Install dependencies
       run: |
-        if [ "${{ matrix.transformers_version }}" = "main" ]; then
-          uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
-        else
-          uv pip uninstall transformers huggingface_hub && uv pip install transformers==${{ matrix.transformers_version }}
-        fi
+        #uv pip uninstall transformers huggingface_hub && uv pip install --prerelease allow -U transformers@git+https://github.com/huggingface/transformers.git
+        uv pip uninstall transformers huggingface_hub && uv pip install transformers==4.57.1
         uv pip install -e ".[quality,training]"
 
     - name: Environment
@@ -307,6 +289,6 @@ jobs:
       if: ${{ always() }}
       uses: actions/upload-artifact@v6
       with:
-        name: examples_transformers_${{ matrix.transformers_version }}_test_reports
+        name: examples_test_reports
         path: reports
 

examples/custom_diffusion/test_custom_diffusion.py

@@ -17,9 +17,6 @@ import logging
 import os
 import sys
 import tempfile
-import unittest
-
-from diffusers.utils import is_transformers_version
 
 
 sys.path.append("..")
@@ -33,7 +30,6 @@ stream_handler = logging.StreamHandler(sys.stdout)
 logger.addHandler(stream_handler)
 
 
-@unittest.skipIf(is_transformers_version(">=", "4.57.5"), "Size mismatch")
 class CustomDiffusion(ExamplesTestsAccelerate):
     def test_custom_diffusion(self):
         with tempfile.TemporaryDirectory() as tmpdir:

src/diffusers/hooks/_common.py

@@ -44,7 +44,6 @@ _GO_LC_SUPPORTED_PYTORCH_LAYERS = (
     torch.nn.ConvTranspose2d,
     torch.nn.ConvTranspose3d,
     torch.nn.Linear,
-    torch.nn.Embedding,
     # TODO(aryan): look into torch.nn.LayerNorm, torch.nn.GroupNorm later, seems to be causing some issues with CogVideoX
     # because of double invocation of the same norm layer in CogVideoXLayerNorm
 )

src/diffusers/loaders/textual_inversion.py

@@ -21,12 +21,7 @@ from tokenizers import Tokenizer as TokenizerFast
 from torch import nn
 
 from ..models.modeling_utils import load_state_dict
-from ..utils import (
-    _get_model_file,
-    is_accelerate_available,
-    is_transformers_available,
-    logging,
-)
+from ..utils import _get_model_file, is_accelerate_available, is_transformers_available, logging
 
 
 if is_transformers_available():

src/diffusers/pipelines/cosmos/pipeline_cosmos2_5_predict.py

@@ -287,9 +287,6 @@ class Cosmos2_5_PredictBasePipeline(DiffusionPipeline):
                 truncation=True,
                 padding="max_length",
             )
-            input_ids = (
-                input_ids["input_ids"] if not isinstance(input_ids, list) and "input_ids" in input_ids else input_ids
-            )
             input_ids = torch.LongTensor(input_ids)
             input_ids_batch.append(input_ids)
 

src/diffusers/pipelines/kandinsky/text_encoder.py

@@ -20,8 +20,6 @@ class MultilingualCLIP(PreTrainedModel):
         self.LinearTransformation = torch.nn.Linear(
             in_features=config.transformerDimensions, out_features=config.numDims
         )
-        if hasattr(self, "post_init"):
-            self.post_init()
 
     def forward(self, input_ids, attention_mask):
         embs = self.transformer(input_ids=input_ids, attention_mask=attention_mask)[0]

src/diffusers/pipelines/kolors/text_encoder.py

@@ -782,9 +782,6 @@ class ChatGLMModel(ChatGLMPreTrainedModel):
             self.prefix_encoder = PrefixEncoder(config)
             self.dropout = torch.nn.Dropout(0.1)
 
-        if hasattr(self, "post_init"):
-            self.post_init()
-
     def get_input_embeddings(self):
         return self.embedding.word_embeddings
 
@@ -814,7 +811,7 @@ class ChatGLMModel(ChatGLMPreTrainedModel):
         output_hidden_states = (
             output_hidden_states if output_hidden_states is not None else self.config.output_hidden_states
         )
-        use_cache = use_cache if use_cache is not None else getattr(self.config, "use_cache", None)
+        use_cache = use_cache if use_cache is not None else self.config.use_cache
         return_dict = return_dict if return_dict is not None else self.config.use_return_dict
 
         batch_size, seq_length = input_ids.shape

src/diffusers/pipelines/pipeline_utils.py

@@ -340,7 +340,6 @@ class DiffusionPipeline(ConfigMixin, PushToHubMixin):
             save_method_accept_safe = "safe_serialization" in save_method_signature.parameters
             save_method_accept_variant = "variant" in save_method_signature.parameters
             save_method_accept_max_shard_size = "max_shard_size" in save_method_signature.parameters
-            save_method_accept_peft_format = "save_peft_format" in save_method_signature.parameters
 
             save_kwargs = {}
             if save_method_accept_safe:
@@ -350,11 +349,6 @@ class DiffusionPipeline(ConfigMixin, PushToHubMixin):
             if save_method_accept_max_shard_size and max_shard_size is not None:
                 # max_shard_size is expected to not be None in ModelMixin
                 save_kwargs["max_shard_size"] = max_shard_size
-            if save_method_accept_peft_format:
-                # Set save_peft_format=False for transformers>=5.0.0 compatibility
-                # In transformers 5.0.0+, the default save_peft_format=True adds "base_model.model" prefix
-                # to adapter keys, but from_pretrained expects keys without this prefix
-                save_kwargs["save_peft_format"] = False
 
             save_method(os.path.join(save_directory, pipeline_component_name), **save_kwargs)
 

tests/models/test_models_auto.py

@@ -20,9 +20,7 @@ class TestAutoModel(unittest.TestCase):
         side_effect=[EnvironmentError("File not found"), {"model_type": "clip_text_model"}],
     )
     def test_load_from_config_transformers_with_subfolder(self, mock_load_config):
-        model = AutoModel.from_pretrained(
-            "hf-internal-testing/tiny-stable-diffusion-torch", subfolder="text_encoder", use_safetensors=False
-        )
+        model = AutoModel.from_pretrained("hf-internal-testing/tiny-stable-diffusion-torch", subfolder="text_encoder")
         assert isinstance(model, CLIPTextModel)
 
     def test_load_from_config_without_subfolder(self):
@@ -30,7 +28,5 @@ class TestAutoModel(unittest.TestCase):
         assert isinstance(model, LongformerModel)
 
     def test_load_from_model_index(self):
-        model = AutoModel.from_pretrained(
-            "hf-internal-testing/tiny-stable-diffusion-torch", subfolder="text_encoder", use_safetensors=False
-        )
+        model = AutoModel.from_pretrained("hf-internal-testing/tiny-stable-diffusion-torch", subfolder="text_encoder")
         assert isinstance(model, CLIPTextModel)

tests/models/transformers/test_models_transformer_wan.py

@@ -12,57 +12,52 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-
 import torch
 
 from diffusers import WanTransformer3DModel
+from diffusers.utils.torch_utils import randn_tensor
 
-from ...testing_utils import (
-    enable_full_determinism,
-    torch_device,
+from ...testing_utils import enable_full_determinism, torch_device
+from ..testing_utils import (
+    AttentionTesterMixin,
+    BaseModelTesterConfig,
+    BitsAndBytesTesterMixin,
+    GGUFCompileTesterMixin,
+    GGUFTesterMixin,
+    MemoryTesterMixin,
+    ModelTesterMixin,
+    TorchAoTesterMixin,
+    TorchCompileTesterMixin,
+    TrainingTesterMixin,
 )
-from ..test_modeling_common import ModelTesterMixin, TorchCompileTesterMixin
 
 
 enable_full_determinism()
 
 
-class WanTransformer3DTests(ModelTesterMixin, unittest.TestCase):
-    model_class = WanTransformer3DModel
-    main_input_name = "hidden_states"
-    uses_custom_attn_processor = True
+class WanTransformer3DTesterConfig(BaseModelTesterConfig):
+    @property
+    def model_class(self):
+        return WanTransformer3DModel
 
     @property
-    def dummy_input(self):
-        batch_size = 1
-        num_channels = 4
-        num_frames = 2
-        height = 16
-        width = 16
-        text_encoder_embedding_dim = 16
-        sequence_length = 12
+    def output_shape(self) -> tuple[int, ...]:
+        return (4, 2, 16, 16)
 
-        hidden_states = torch.randn((batch_size, num_channels, num_frames, height, width)).to(torch_device)
-        timestep = torch.randint(0, 1000, size=(batch_size,)).to(torch_device)
-        encoder_hidden_states = torch.randn((batch_size, sequence_length, text_encoder_embedding_dim)).to(torch_device)
+    @property
+    def input_shape(self) -> tuple[int, ...]:
+        return (4, 2, 16, 16)
 
+    @property
+    def main_input_name(self) -> str:
+        return "hidden_states"
+
+    @property
+    def generator(self):
+        return torch.Generator("cpu").manual_seed(0)
+
+    def get_init_dict(self) -> dict[str, int | list[int] | tuple | str | bool]:
         return {
-            "hidden_states": hidden_states,
-            "encoder_hidden_states": encoder_hidden_states,
-            "timestep": timestep,
-        }
-
-    @property
-    def input_shape(self):
-        return (4, 1, 16, 16)
-
-    @property
-    def output_shape(self):
-        return (4, 1, 16, 16)
-
-    def prepare_init_args_and_inputs_for_common(self):
-        init_dict = {
             "patch_size": (1, 2, 2),
             "num_attention_heads": 2,
             "attention_head_dim": 12,
@@ -76,16 +71,118 @@ class WanTransformer3DTests(ModelTesterMixin, unittest.TestCase):
             "qk_norm": "rms_norm_across_heads",
             "rope_max_seq_len": 32,
         }
-        inputs_dict = self.dummy_input
-        return init_dict, inputs_dict
+
+    def get_dummy_inputs(self) -> dict[str, torch.Tensor]:
+        batch_size = 1
+        num_channels = 4
+        num_frames = 2
+        height = 16
+        width = 16
+        text_encoder_embedding_dim = 16
+        sequence_length = 12
+
+        return {
+            "hidden_states": randn_tensor(
+                (batch_size, num_channels, num_frames, height, width),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (batch_size, sequence_length, text_encoder_embedding_dim),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "timestep": torch.randint(0, 1000, size=(batch_size,), generator=self.generator).to(torch_device),
+        }
+
+
+class TestWanTransformer3D(WanTransformer3DTesterConfig, ModelTesterMixin):
+    """Core model tests for Wan Transformer 3D."""
+
+
+class TestWanTransformer3DMemory(WanTransformer3DTesterConfig, MemoryTesterMixin):
+    """Memory optimization tests for Wan Transformer 3D."""
+
+
+class TestWanTransformer3DTraining(WanTransformer3DTesterConfig, TrainingTesterMixin):
+    """Training tests for Wan Transformer 3D."""
 
     def test_gradient_checkpointing_is_applied(self):
         expected_set = {"WanTransformer3DModel"}
         super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
 
 
-class WanTransformerCompileTests(TorchCompileTesterMixin, unittest.TestCase):
-    model_class = WanTransformer3DModel
+class TestWanTransformer3DAttention(WanTransformer3DTesterConfig, AttentionTesterMixin):
+    """Attention processor tests for Wan Transformer 3D."""
 
-    def prepare_init_args_and_inputs_for_common(self):
-        return WanTransformer3DTests().prepare_init_args_and_inputs_for_common()
+
+class TestWanTransformer3DCompile(WanTransformer3DTesterConfig, TorchCompileTesterMixin):
+    """Torch compile tests for Wan Transformer 3D."""
+
+
+class TestWanTransformer3DBitsAndBytes(WanTransformer3DTesterConfig, BitsAndBytesTesterMixin):
+    """BitsAndBytes quantization tests for Wan Transformer 3D."""
+
+
+class TestWanTransformer3DTorchAo(WanTransformer3DTesterConfig, TorchAoTesterMixin):
+    """TorchAO quantization tests for Wan Transformer 3D."""
+
+
+class TestWanTransformer3DGGUF(WanTransformer3DTesterConfig, GGUFTesterMixin):
+    """GGUF quantization tests for Wan Transformer 3D."""
+
+    @property
+    def gguf_filename(self):
+        return "https://huggingface.co/QuantStack/Wan2.2-I2V-A14B-GGUF/blob/main/LowNoise/Wan2.2-I2V-A14B-LowNoise-Q2_K.gguf"
+
+    @property
+    def torch_dtype(self):
+        return torch.bfloat16
+
+    def get_dummy_inputs(self):
+        """Override to provide inputs matching the real Wan I2V model dimensions.
+
+        Wan 2.2 I2V: in_channels=36, text_dim=4096, image_dim=1280
+        """
+        return {
+            "hidden_states": randn_tensor(
+                (1, 36, 2, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (1, 512, 4096), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states_image": randn_tensor(
+                (1, 257, 1280), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "timestep": torch.tensor([1.0]).to(torch_device, self.torch_dtype),
+        }
+
+
+class TestWanTransformer3DGGUFCompile(WanTransformer3DTesterConfig, GGUFCompileTesterMixin):
+    """GGUF + compile tests for Wan Transformer 3D."""
+
+    @property
+    def gguf_filename(self):
+        return "https://huggingface.co/QuantStack/Wan2.2-I2V-A14B-GGUF/blob/main/LowNoise/Wan2.2-I2V-A14B-LowNoise-Q2_K.gguf"
+
+    @property
+    def torch_dtype(self):
+        return torch.bfloat16
+
+    def get_dummy_inputs(self):
+        """Override to provide inputs matching the real Wan I2V model dimensions.
+
+        Wan 2.2 I2V: in_channels=36, text_dim=4096, image_dim=1280
+        """
+        return {
+            "hidden_states": randn_tensor(
+                (1, 36, 2, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (1, 512, 4096), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states_image": randn_tensor(
+                (1, 257, 1280), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "timestep": torch.tensor([1.0]).to(torch_device, self.torch_dtype),
+        }

tests/models/transformers/test_models_transformer_wan_animate.py

@@ -12,76 +12,57 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import unittest
-
 import torch
 
 from diffusers import WanAnimateTransformer3DModel
+from diffusers.utils.torch_utils import randn_tensor
 
-from ...testing_utils import (
-    enable_full_determinism,
-    torch_device,
+from ...testing_utils import enable_full_determinism, torch_device
+from ..testing_utils import (
+    AttentionTesterMixin,
+    BaseModelTesterConfig,
+    BitsAndBytesTesterMixin,
+    GGUFCompileTesterMixin,
+    GGUFTesterMixin,
+    MemoryTesterMixin,
+    ModelTesterMixin,
+    TorchAoTesterMixin,
+    TorchCompileTesterMixin,
+    TrainingTesterMixin,
 )
-from ..test_modeling_common import ModelTesterMixin, TorchCompileTesterMixin
 
 
 enable_full_determinism()
 
 
-class WanAnimateTransformer3DTests(ModelTesterMixin, unittest.TestCase):
-    model_class = WanAnimateTransformer3DModel
-    main_input_name = "hidden_states"
-    uses_custom_attn_processor = True
+class WanAnimateTransformer3DTesterConfig(BaseModelTesterConfig):
+    @property
+    def model_class(self):
+        return WanAnimateTransformer3DModel
 
     @property
-    def dummy_input(self):
-        batch_size = 1
-        num_channels = 4
-        num_frames = 20  # To make the shapes work out; for complicated reasons we want 21 to divide num_frames + 1
-        height = 16
-        width = 16
-        text_encoder_embedding_dim = 16
-        sequence_length = 12
-
-        clip_seq_len = 12
-        clip_dim = 16
-
-        inference_segment_length = 77  # The inference segment length in the full Wan2.2-Animate-14B model
-        face_height = 16  # Should be square and match `motion_encoder_size` below
-        face_width = 16
-
-        hidden_states = torch.randn((batch_size, 2 * num_channels + 4, num_frames + 1, height, width)).to(torch_device)
-        timestep = torch.randint(0, 1000, size=(batch_size,)).to(torch_device)
-        encoder_hidden_states = torch.randn((batch_size, sequence_length, text_encoder_embedding_dim)).to(torch_device)
-        clip_ref_features = torch.randn((batch_size, clip_seq_len, clip_dim)).to(torch_device)
-        pose_latents = torch.randn((batch_size, num_channels, num_frames, height, width)).to(torch_device)
-        face_pixel_values = torch.randn((batch_size, 3, inference_segment_length, face_height, face_width)).to(
-            torch_device
-        )
-
-        return {
-            "hidden_states": hidden_states,
-            "timestep": timestep,
-            "encoder_hidden_states": encoder_hidden_states,
-            "encoder_hidden_states_image": clip_ref_features,
-            "pose_hidden_states": pose_latents,
-            "face_pixel_values": face_pixel_values,
-        }
+    def output_shape(self) -> tuple[int, ...]:
+        # Output has fewer channels than input (4 vs 12)
+        return (4, 21, 16, 16)
 
     @property
-    def input_shape(self):
-        return (12, 1, 16, 16)
+    def input_shape(self) -> tuple[int, ...]:
+        return (12, 21, 16, 16)
 
     @property
-    def output_shape(self):
-        return (4, 1, 16, 16)
+    def main_input_name(self) -> str:
+        return "hidden_states"
 
-    def prepare_init_args_and_inputs_for_common(self):
+    @property
+    def generator(self):
+        return torch.Generator("cpu").manual_seed(0)
+
+    def get_init_dict(self) -> dict[str, int | list[int] | tuple | str | bool | float | dict]:
         # Use custom channel sizes since the default Wan Animate channel sizes will cause the motion encoder to
         # contain the vast majority of the parameters in the test model
         channel_sizes = {"4": 16, "8": 16, "16": 16}
 
-        init_dict = {
+        return {
             "patch_size": (1, 2, 2),
             "num_attention_heads": 2,
             "attention_head_dim": 12,
@@ -105,22 +86,158 @@ class WanAnimateTransformer3DTests(ModelTesterMixin, unittest.TestCase):
             "face_encoder_num_heads": 2,
             "inject_face_latents_blocks": 2,
         }
-        inputs_dict = self.dummy_input
-        return init_dict, inputs_dict
+
+    def get_dummy_inputs(self) -> dict[str, torch.Tensor]:
+        batch_size = 1
+        num_channels = 4
+        num_frames = 20  # To make the shapes work out; for complicated reasons we want 21 to divide num_frames + 1
+        height = 16
+        width = 16
+        text_encoder_embedding_dim = 16
+        sequence_length = 12
+
+        clip_seq_len = 12
+        clip_dim = 16
+
+        inference_segment_length = 77  # The inference segment length in the full Wan2.2-Animate-14B model
+        face_height = 16  # Should be square and match `motion_encoder_size`
+        face_width = 16
+
+        return {
+            "hidden_states": randn_tensor(
+                (batch_size, 2 * num_channels + 4, num_frames + 1, height, width),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "timestep": torch.randint(0, 1000, size=(batch_size,), generator=self.generator).to(torch_device),
+            "encoder_hidden_states": randn_tensor(
+                (batch_size, sequence_length, text_encoder_embedding_dim),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "encoder_hidden_states_image": randn_tensor(
+                (batch_size, clip_seq_len, clip_dim),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "pose_hidden_states": randn_tensor(
+                (batch_size, num_channels, num_frames, height, width),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "face_pixel_values": randn_tensor(
+                (batch_size, 3, inference_segment_length, face_height, face_width),
+                generator=self.generator,
+                device=torch_device,
+            ),
+        }
+
+
+class TestWanAnimateTransformer3D(WanAnimateTransformer3DTesterConfig, ModelTesterMixin):
+    """Core model tests for Wan Animate Transformer 3D."""
+
+    def test_output(self):
+        # Override test_output because the transformer output is expected to have less channels
+        # than the main transformer input.
+        expected_output_shape = (1, 4, 21, 16, 16)
+        super().test_output(expected_output_shape=expected_output_shape)
+
+
+class TestWanAnimateTransformer3DMemory(WanAnimateTransformer3DTesterConfig, MemoryTesterMixin):
+    """Memory optimization tests for Wan Animate Transformer 3D."""
+
+
+class TestWanAnimateTransformer3DTraining(WanAnimateTransformer3DTesterConfig, TrainingTesterMixin):
+    """Training tests for Wan Animate Transformer 3D."""
 
     def test_gradient_checkpointing_is_applied(self):
         expected_set = {"WanAnimateTransformer3DModel"}
         super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
 
-    # Override test_output because the transformer output is expected to have less channels than the main transformer
-    # input.
-    def test_output(self):
-        expected_output_shape = (1, 4, 21, 16, 16)
-        super().test_output(expected_output_shape=expected_output_shape)
+
+class TestWanAnimateTransformer3DAttention(WanAnimateTransformer3DTesterConfig, AttentionTesterMixin):
+    """Attention processor tests for Wan Animate Transformer 3D."""
 
 
-class WanAnimateTransformerCompileTests(TorchCompileTesterMixin, unittest.TestCase):
-    model_class = WanAnimateTransformer3DModel
+class TestWanAnimateTransformer3DCompile(WanAnimateTransformer3DTesterConfig, TorchCompileTesterMixin):
+    """Torch compile tests for Wan Animate Transformer 3D."""
 
-    def prepare_init_args_and_inputs_for_common(self):
-        return WanAnimateTransformer3DTests().prepare_init_args_and_inputs_for_common()
+
+class TestWanAnimateTransformer3DBitsAndBytes(WanAnimateTransformer3DTesterConfig, BitsAndBytesTesterMixin):
+    """BitsAndBytes quantization tests for Wan Animate Transformer 3D."""
+
+
+class TestWanAnimateTransformer3DTorchAo(WanAnimateTransformer3DTesterConfig, TorchAoTesterMixin):
+    """TorchAO quantization tests for Wan Animate Transformer 3D."""
+
+
+class TestWanAnimateTransformer3DGGUF(WanAnimateTransformer3DTesterConfig, GGUFTesterMixin):
+    """GGUF quantization tests for Wan Animate Transformer 3D."""
+
+    @property
+    def gguf_filename(self):
+        return "https://huggingface.co/QuantStack/Wan2.2-Animate-14B-GGUF/blob/main/Wan2.2-Animate-14B-Q2_K.gguf"
+
+    @property
+    def torch_dtype(self):
+        return torch.bfloat16
+
+    def get_dummy_inputs(self):
+        """Override to provide inputs matching the real Wan Animate model dimensions.
+
+        Wan 2.2 Animate: in_channels=36 (2*16+4), text_dim=4096, image_dim=1280
+        """
+        return {
+            "hidden_states": randn_tensor(
+                (1, 36, 21, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (1, 512, 4096), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states_image": randn_tensor(
+                (1, 257, 1280), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "pose_hidden_states": randn_tensor(
+                (1, 16, 20, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "face_pixel_values": randn_tensor(
+                (1, 3, 77, 512, 512), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "timestep": torch.tensor([1.0]).to(torch_device, self.torch_dtype),
+        }
+
+
+class TestWanAnimateTransformer3DGGUFCompile(WanAnimateTransformer3DTesterConfig, GGUFCompileTesterMixin):
+    """GGUF + compile tests for Wan Animate Transformer 3D."""
+
+    @property
+    def gguf_filename(self):
+        return "https://huggingface.co/QuantStack/Wan2.2-Animate-14B-GGUF/blob/main/Wan2.2-Animate-14B-Q2_K.gguf"
+
+    @property
+    def torch_dtype(self):
+        return torch.bfloat16
+
+    def get_dummy_inputs(self):
+        """Override to provide inputs matching the real Wan Animate model dimensions.
+
+        Wan 2.2 Animate: in_channels=36 (2*16+4), text_dim=4096, image_dim=1280
+        """
+        return {
+            "hidden_states": randn_tensor(
+                (1, 36, 21, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (1, 512, 4096), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states_image": randn_tensor(
+                (1, 257, 1280), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "pose_hidden_states": randn_tensor(
+                (1, 16, 20, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "face_pixel_values": randn_tensor(
+                (1, 3, 77, 512, 512), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "timestep": torch.tensor([1.0]).to(torch_device, self.torch_dtype),
+        }

tests/models/transformers/test_models_transformer_wan_vace.py

@@ -0,0 +1,198 @@
+# Copyright 2025 HuggingFace Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import torch
+
+from diffusers import WanVACETransformer3DModel
+from diffusers.utils.torch_utils import randn_tensor
+
+from ...testing_utils import enable_full_determinism, torch_device
+from ..testing_utils import (
+    AttentionTesterMixin,
+    BaseModelTesterConfig,
+    BitsAndBytesTesterMixin,
+    GGUFCompileTesterMixin,
+    GGUFTesterMixin,
+    MemoryTesterMixin,
+    ModelTesterMixin,
+    TorchAoTesterMixin,
+    TorchCompileTesterMixin,
+    TrainingTesterMixin,
+)
+
+
+enable_full_determinism()
+
+
+class WanVACETransformer3DTesterConfig(BaseModelTesterConfig):
+    @property
+    def model_class(self):
+        return WanVACETransformer3DModel
+
+    @property
+    def output_shape(self) -> tuple[int, ...]:
+        return (16, 2, 16, 16)
+
+    @property
+    def input_shape(self) -> tuple[int, ...]:
+        return (16, 2, 16, 16)
+
+    @property
+    def main_input_name(self) -> str:
+        return "hidden_states"
+
+    @property
+    def generator(self):
+        return torch.Generator("cpu").manual_seed(0)
+
+    def get_init_dict(self) -> dict[str, int | list[int] | tuple | str | bool | None]:
+        return {
+            "patch_size": (1, 2, 2),
+            "num_attention_heads": 2,
+            "attention_head_dim": 12,
+            "in_channels": 16,
+            "out_channels": 16,
+            "text_dim": 32,
+            "freq_dim": 256,
+            "ffn_dim": 32,
+            "num_layers": 4,
+            "cross_attn_norm": True,
+            "qk_norm": "rms_norm_across_heads",
+            "rope_max_seq_len": 32,
+            "vace_layers": [0, 2],
+            "vace_in_channels": 48,  # 3 * in_channels = 3 * 16 = 48
+        }
+
+    def get_dummy_inputs(self) -> dict[str, torch.Tensor]:
+        batch_size = 1
+        num_channels = 16
+        num_frames = 2
+        height = 16
+        width = 16
+        text_encoder_embedding_dim = 32
+        sequence_length = 12
+
+        # VACE requires control_hidden_states with vace_in_channels (3 * in_channels)
+        vace_in_channels = 48
+
+        return {
+            "hidden_states": randn_tensor(
+                (batch_size, num_channels, num_frames, height, width),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (batch_size, sequence_length, text_encoder_embedding_dim),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "control_hidden_states": randn_tensor(
+                (batch_size, vace_in_channels, num_frames, height, width),
+                generator=self.generator,
+                device=torch_device,
+            ),
+            "timestep": torch.randint(0, 1000, size=(batch_size,), generator=self.generator).to(torch_device),
+        }
+
+
+class TestWanVACETransformer3D(WanVACETransformer3DTesterConfig, ModelTesterMixin):
+    """Core model tests for Wan VACE Transformer 3D."""
+
+
+class TestWanVACETransformer3DMemory(WanVACETransformer3DTesterConfig, MemoryTesterMixin):
+    """Memory optimization tests for Wan VACE Transformer 3D."""
+
+
+class TestWanVACETransformer3DTraining(WanVACETransformer3DTesterConfig, TrainingTesterMixin):
+    """Training tests for Wan VACE Transformer 3D."""
+
+    def test_gradient_checkpointing_is_applied(self):
+        expected_set = {"WanVACETransformer3DModel"}
+        super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
+
+
+class TestWanVACETransformer3DAttention(WanVACETransformer3DTesterConfig, AttentionTesterMixin):
+    """Attention processor tests for Wan VACE Transformer 3D."""
+
+
+class TestWanVACETransformer3DCompile(WanVACETransformer3DTesterConfig, TorchCompileTesterMixin):
+    """Torch compile tests for Wan VACE Transformer 3D."""
+
+
+class TestWanVACETransformer3DBitsAndBytes(WanVACETransformer3DTesterConfig, BitsAndBytesTesterMixin):
+    """BitsAndBytes quantization tests for Wan VACE Transformer 3D."""
+
+
+class TestWanVACETransformer3DTorchAo(WanVACETransformer3DTesterConfig, TorchAoTesterMixin):
+    """TorchAO quantization tests for Wan VACE Transformer 3D."""
+
+
+class TestWanVACETransformer3DGGUF(WanVACETransformer3DTesterConfig, GGUFTesterMixin):
+    """GGUF quantization tests for Wan VACE Transformer 3D."""
+
+    @property
+    def gguf_filename(self):
+        return "https://huggingface.co/QuantStack/Wan2.1_14B_VACE-GGUF/blob/main/Wan2.1_14B_VACE-Q3_K_S.gguf"
+
+    @property
+    def torch_dtype(self):
+        return torch.bfloat16
+
+    def get_dummy_inputs(self):
+        """Override to provide inputs matching the real Wan VACE model dimensions.
+
+        Wan 2.1 VACE: in_channels=16, text_dim=4096, vace_in_channels=96
+        """
+        return {
+            "hidden_states": randn_tensor(
+                (1, 16, 2, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (1, 512, 4096), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "control_hidden_states": randn_tensor(
+                (1, 96, 2, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "timestep": torch.tensor([1.0]).to(torch_device, self.torch_dtype),
+        }
+
+
+class TestWanVACETransformer3DGGUFCompile(WanVACETransformer3DTesterConfig, GGUFCompileTesterMixin):
+    """GGUF + compile tests for Wan VACE Transformer 3D."""
+
+    @property
+    def gguf_filename(self):
+        return "https://huggingface.co/QuantStack/Wan2.1_14B_VACE-GGUF/blob/main/Wan2.1_14B_VACE-Q3_K_S.gguf"
+
+    @property
+    def torch_dtype(self):
+        return torch.bfloat16
+
+    def get_dummy_inputs(self):
+        """Override to provide inputs matching the real Wan VACE model dimensions.
+
+        Wan 2.1 VACE: in_channels=16, text_dim=4096, vace_in_channels=96
+        """
+        return {
+            "hidden_states": randn_tensor(
+                (1, 16, 2, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (1, 512, 4096), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "control_hidden_states": randn_tensor(
+                (1, 96, 2, 64, 64), generator=self.generator, device=torch_device, dtype=self.torch_dtype
+            ),
+            "timestep": torch.tensor([1.0]).to(torch_device, self.torch_dtype),
+        }

tests/pipelines/bria/test_pipeline_bria.py

@@ -19,7 +19,7 @@ import unittest
 import numpy as np
 import torch
 from huggingface_hub import hf_hub_download
-from transformers import AutoConfig, T5EncoderModel, T5TokenizerFast
+from transformers import T5EncoderModel, T5TokenizerFast
 
 from diffusers import (
     AutoencoderKL,
@@ -89,8 +89,7 @@ class BriaPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         scheduler = FlowMatchEulerDiscreteScheduler()
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = T5TokenizerFast.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/chroma/test_pipeline_chroma.py

@@ -2,7 +2,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, ChromaPipeline, ChromaTransformer2DModel, FlowMatchEulerDiscreteScheduler
 
@@ -41,8 +41,7 @@ class ChromaPipelineFastTests(
         )
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 

tests/pipelines/chroma/test_pipeline_chroma_img2img.py

@@ -3,7 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, ChromaImg2ImgPipeline, ChromaTransformer2DModel, FlowMatchEulerDiscreteScheduler
 
@@ -42,8 +42,7 @@ class ChromaImg2ImgPipelineFastTests(
         )
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 

tests/pipelines/chronoedit/test_chronoedit.py

@@ -17,7 +17,6 @@ import unittest
 import torch
 from PIL import Image
 from transformers import (
-    AutoConfig,
     AutoTokenizer,
     CLIPImageProcessor,
     CLIPVisionConfig,
@@ -72,8 +71,7 @@ class ChronoEditPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         torch.manual_seed(0)
         # TODO: impl FlowDPMSolverMultistepScheduler
         scheduler = FlowMatchEulerDiscreteScheduler(shift=7.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/cogvideo/test_cogvideox.py

@@ -18,7 +18,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCogVideoX, CogVideoXPipeline, CogVideoXTransformer3DModel, DDIMScheduler
 
@@ -117,8 +117,7 @@ class CogVideoXPipelineFastTests(
 
         torch.manual_seed(0)
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cogvideo/test_cogvideox_fun_control.py

@@ -18,7 +18,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCogVideoX, CogVideoXFunControlPipeline, CogVideoXTransformer3DModel, DDIMScheduler
 
@@ -104,8 +104,7 @@ class CogVideoXFunControlPipelineFastTests(PipelineTesterMixin, unittest.TestCas
 
         torch.manual_seed(0)
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cogvideo/test_cogvideox_image2video.py

@@ -19,7 +19,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCogVideoX, CogVideoXImageToVideoPipeline, CogVideoXTransformer3DModel, DDIMScheduler
 from diffusers.utils import load_image
@@ -113,8 +113,7 @@ class CogVideoXImageToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestC
 
         torch.manual_seed(0)
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cogvideo/test_cogvideox_video2video.py

@@ -18,7 +18,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCogVideoX, CogVideoXTransformer3DModel, CogVideoXVideoToVideoPipeline, DDIMScheduler
 
@@ -99,8 +99,7 @@ class CogVideoXVideoToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestC
 
         torch.manual_seed(0)
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cogview3/test_cogview3plus.py

@@ -18,7 +18,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, CogVideoXDDIMScheduler, CogView3PlusPipeline, CogView3PlusTransformer2DModel
 
@@ -89,8 +89,7 @@ class CogView3PlusPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = CogVideoXDDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cogview4/test_cogview4.py

@@ -108,7 +108,7 @@ class CogView4PipelineFastTests(PipelineTesterMixin, unittest.TestCase):
             generator = torch.Generator(device=device).manual_seed(seed)
         inputs = {
             "prompt": "dance monkey",
-            "negative_prompt": "bad",
+            "negative_prompt": "",
             "generator": generator,
             "num_inference_steps": 2,
             "guidance_scale": 6.0,

tests/pipelines/consisid/test_consisid.py

@@ -19,7 +19,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCogVideoX, ConsisIDPipeline, ConsisIDTransformer3DModel, DDIMScheduler
 from diffusers.utils import load_image
@@ -122,8 +122,7 @@ class ConsisIDPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/controlnet_flux/test_controlnet_flux.py

@@ -19,7 +19,7 @@ import unittest
 import numpy as np
 import torch
 from huggingface_hub import hf_hub_download
-from transformers import AutoConfig, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel, T5TokenizerFast
+from transformers import CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel, T5TokenizerFast
 
 from diffusers import (
     AutoencoderKL,
@@ -97,8 +97,7 @@ class FluxControlNetPipelineFastTests(unittest.TestCase, PipelineTesterMixin, Fl
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = T5TokenizerFast.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/controlnet_flux/test_controlnet_flux_img2img.py

@@ -2,7 +2,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -13,7 +13,9 @@ from diffusers import (
 )
 from diffusers.utils.torch_utils import randn_tensor
 
-from ...testing_utils import torch_device
+from ...testing_utils import (
+    torch_device,
+)
 from ..test_pipelines_common import PipelineTesterMixin, check_qkv_fused_layers_exist
 
 
@@ -68,8 +70,7 @@ class FluxControlNetImg2ImgPipelineFastTests(unittest.TestCase, PipelineTesterMi
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/controlnet_flux/test_controlnet_flux_inpaint.py

@@ -3,7 +3,15 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+
+# torch_device,  # {{ edit_1 }} Removed unused import
+from transformers import (
+    AutoTokenizer,
+    CLIPTextConfig,
+    CLIPTextModel,
+    CLIPTokenizer,
+    T5EncoderModel,
+)
 
 from diffusers import (
     AutoencoderKL,
@@ -14,7 +22,11 @@ from diffusers import (
 )
 from diffusers.utils.torch_utils import randn_tensor
 
-from ...testing_utils import enable_full_determinism, floats_tensor, torch_device
+from ...testing_utils import (
+    enable_full_determinism,
+    floats_tensor,
+    torch_device,
+)
 from ..test_pipelines_common import PipelineTesterMixin
 
 
@@ -73,8 +85,7 @@ class FluxControlNetInpaintPipelineTests(unittest.TestCase, PipelineTesterMixin)
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/controlnet_hunyuandit/test_controlnet_hunyuandit.py

@@ -18,7 +18,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, BertModel, T5EncoderModel
+from transformers import AutoTokenizer, BertModel, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -96,10 +96,7 @@ class HunyuanDiTControlNetPipelineFastTests(unittest.TestCase, PipelineTesterMix
         scheduler = DDPMScheduler()
         text_encoder = BertModel.from_pretrained("hf-internal-testing/tiny-random-BertModel")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-BertModel")
-
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/controlnet_sd3/test_controlnet_inpaint_sd3.py

@@ -17,14 +17,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -35,7 +28,10 @@ from diffusers import (
 from diffusers.models import SD3ControlNetModel
 from diffusers.utils.torch_utils import randn_tensor
 
-from ...testing_utils import enable_full_determinism, torch_device
+from ...testing_utils import (
+    enable_full_determinism,
+    torch_device,
+)
 from ..test_pipelines_common import PipelineTesterMixin
 
 
@@ -107,8 +103,7 @@ class StableDiffusion3ControlInpaintNetPipelineFastTests(unittest.TestCase, Pipe
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/controlnet_sd3/test_controlnet_sd3.py

@@ -19,14 +19,7 @@ from typing import Optional
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -125,8 +118,7 @@ class StableDiffusion3ControlNetPipelineFastTests(unittest.TestCase, PipelineTes
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/cosmos/test_cosmos.py

@@ -20,7 +20,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCosmos, CosmosTextToWorldPipeline, CosmosTransformer3DModel, EDMEulerScheduler
 
@@ -107,8 +107,7 @@ class CosmosTextToWorldPipelineFastTests(PipelineTesterMixin, unittest.TestCase)
             rho=7.0,
             final_sigmas_type="sigma_min",
         )
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cosmos/test_cosmos2_text2image.py

@@ -20,7 +20,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLWan,
@@ -95,8 +95,7 @@ class Cosmos2TextToImagePipelineFastTests(PipelineTesterMixin, unittest.TestCase
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler(use_karras_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cosmos/test_cosmos2_video2world.py

@@ -21,7 +21,7 @@ import unittest
 import numpy as np
 import PIL.Image
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLWan,
@@ -96,8 +96,7 @@ class Cosmos2VideoToWorldPipelineFastTests(PipelineTesterMixin, unittest.TestCas
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler(use_karras_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/cosmos/test_cosmos_video2world.py

@@ -21,7 +21,7 @@ import unittest
 import numpy as np
 import PIL.Image
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLCosmos, CosmosTransformer3DModel, CosmosVideoToWorldPipeline, EDMEulerScheduler
 
@@ -108,8 +108,7 @@ class CosmosVideoToWorldPipelineFastTests(PipelineTesterMixin, unittest.TestCase
             rho=7.0,
             final_sigmas_type="sigma_min",
         )
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/deepfloyd_if/__init__.py

@@ -2,7 +2,7 @@ import tempfile
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import DDPMScheduler, UNet2DConditionModel
 from diffusers.models.attention_processor import AttnAddedKVProcessor
@@ -18,8 +18,7 @@ from ..test_pipelines_common import to_np
 class IFPipelineTesterMixin:
     def _get_dummy_components(self):
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
@@ -76,8 +75,7 @@ class IFPipelineTesterMixin:
 
     def _get_superresolution_dummy_components(self):
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/deepfloyd_if/test_if.py

@@ -18,7 +18,9 @@ import unittest
 
 import torch
 
-from diffusers import IFPipeline
+from diffusers import (
+    IFPipeline,
+)
 from diffusers.models.attention_processor import AttnAddedKVProcessor
 from diffusers.utils.import_utils import is_xformers_available
 

tests/pipelines/flux/test_pipeline_flux.py

@@ -4,7 +4,7 @@ import unittest
 import numpy as np
 import torch
 from huggingface_hub import hf_hub_download
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -91,8 +91,7 @@ class FluxPipelineFastTests(
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_control.py

@@ -3,7 +3,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, FluxControlPipeline, FluxTransformer2DModel
 
@@ -53,8 +53,7 @@ class FluxControlPipelineFastTests(unittest.TestCase, PipelineTesterMixin):
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_control_img2img.py

@@ -3,7 +3,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -57,8 +57,7 @@ class FluxControlImg2ImgPipelineFastTests(unittest.TestCase, PipelineTesterMixin
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_control_inpaint.py

@@ -3,7 +3,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -58,8 +58,7 @@ class FluxControlInpaintPipelineFastTests(unittest.TestCase, PipelineTesterMixin
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_fill.py

@@ -3,7 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, FluxFillPipeline, FluxTransformer2DModel
 
@@ -58,8 +58,7 @@ class FluxFillPipelineFastTests(unittest.TestCase, PipelineTesterMixin):
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_img2img.py

@@ -3,7 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, FluxImg2ImgPipeline, FluxTransformer2DModel
 
@@ -55,8 +55,7 @@ class FluxImg2ImgPipelineFastTests(unittest.TestCase, PipelineTesterMixin, FluxI
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_inpaint.py

@@ -3,7 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, FluxInpaintPipeline, FluxTransformer2DModel
 
@@ -55,8 +55,7 @@ class FluxInpaintPipelineFastTests(unittest.TestCase, PipelineTesterMixin, FluxI
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_kontext.py

@@ -3,7 +3,7 @@ import unittest
 import numpy as np
 import PIL.Image
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -79,8 +79,7 @@ class FluxKontextPipelineFastTests(
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/flux/test_pipeline_flux_kontext_inpaint.py

@@ -3,7 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -79,8 +79,7 @@ class FluxKontextInpaintPipelineFastTests(
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/glm_image/test_glm_image.py

@@ -16,7 +16,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, GlmImagePipeline, GlmImageTransformer2DModel
 from diffusers.utils import is_transformers_version
@@ -57,8 +57,7 @@ class GlmImagePipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
     def get_dummy_components(self):
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         glm_config = GlmImageConfig(

tests/pipelines/hidream_image/test_pipeline_hidream.py

@@ -18,7 +18,6 @@ import unittest
 import numpy as np
 import torch
 from transformers import (
-    AutoConfig,
     AutoTokenizer,
     CLIPTextConfig,
     CLIPTextModelWithProjection,
@@ -95,8 +94,7 @@ class HiDreamImagePipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
         text_encoder_4 = LlamaForCausalLM.from_pretrained("hf-internal-testing/tiny-random-LlamaForCausalLM")
@@ -151,7 +149,7 @@ class HiDreamImagePipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         self.assertEqual(generated_image.shape, (128, 128, 3))
 
         # fmt: off
-        expected_slice = np.array([0.4501, 0.5256, 0.4207, 0.5783, 0.4842, 0.4833, 0.4441, 0.5112, 0.6587, 0.3169, 0.7308, 0.5927, 0.6251, 0.5509, 0.5355, 0.5969])
+        expected_slice = np.array([0.4507, 0.5256, 0.4205, 0.5791, 0.4848, 0.4831, 0.4443, 0.5107, 0.6586, 0.3163, 0.7318, 0.5933, 0.6252, 0.5512, 0.5357, 0.5983])
         # fmt: on
 
         generated_slice = generated_image.flatten()

tests/pipelines/hunyuan_video/test_hunyuan_image2video.py

@@ -233,7 +233,7 @@ class HunyuanVideoImageToVideoPipelineFastTests(
         self.assertEqual(generated_video.shape, (5, 3, 16, 16))
 
         # fmt: off
-        expected_slice = torch.tensor([0.4441, 0.4790, 0.4485, 0.5748, 0.3539, 0.1553, 0.2707, 0.3594, 0.5331, 0.6645, 0.6799, 0.5257, 0.5092, 0.3450, 0.4276, 0.4127])
+        expected_slice = torch.tensor([0.444, 0.479, 0.4485, 0.5752, 0.3539, 0.1548, 0.2706, 0.3593, 0.5323, 0.6635, 0.6795, 0.5255, 0.5091, 0.345, 0.4276, 0.4128])
         # fmt: on
 
         generated_slice = generated_video.flatten()

tests/pipelines/hunyuan_video1_5/test_hunyuan_1_5.py

@@ -15,14 +15,7 @@
 import unittest
 
 import torch
-from transformers import (
-    AutoConfig,
-    ByT5Tokenizer,
-    Qwen2_5_VLTextConfig,
-    Qwen2_5_VLTextModel,
-    Qwen2Tokenizer,
-    T5EncoderModel,
-)
+from transformers import ByT5Tokenizer, Qwen2_5_VLTextConfig, Qwen2_5_VLTextModel, Qwen2Tokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLHunyuanVideo15,
@@ -121,8 +114,7 @@ class HunyuanVideo15PipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         tokenizer = Qwen2Tokenizer.from_pretrained("hf-internal-testing/tiny-random-Qwen2VLForConditionalGeneration")
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer_2 = ByT5Tokenizer()
 
         guider = ClassifierFreeGuidance(guidance_scale=1.0)

tests/pipelines/hunyuandit/test_hunyuan_dit.py

@@ -19,7 +19,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, BertModel, T5EncoderModel
+from transformers import AutoTokenizer, BertModel, T5EncoderModel
 
 from diffusers import AutoencoderKL, DDPMScheduler, HunyuanDiT2DModel, HunyuanDiTPipeline
 
@@ -74,9 +74,7 @@ class HunyuanDiTPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         scheduler = DDPMScheduler()
         text_encoder = BertModel.from_pretrained("hf-internal-testing/tiny-random-BertModel")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-BertModel")
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/kandinsky3/test_kandinsky3.py

@@ -19,7 +19,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoPipelineForImage2Image,
@@ -108,8 +108,7 @@ class Kandinsky3PipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         torch.manual_seed(0)
         movq = self.dummy_movq
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/kandinsky3/test_kandinsky3_img2img.py

@@ -20,7 +20,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoPipelineForImage2Image,
@@ -119,8 +119,7 @@ class Kandinsky3Img2ImgPipelineFastTests(PipelineTesterMixin, unittest.TestCase)
         torch.manual_seed(0)
         movq = self.dummy_movq
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/latte/test_latte.py

@@ -20,7 +20,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -109,8 +109,7 @@ class LattePipelineFastTests(
         vae = AutoencoderKL()
 
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 

tests/pipelines/ltx/test_ltx.py

@@ -17,7 +17,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLLTXVideo, FlowMatchEulerDiscreteScheduler, LTXPipeline, LTXVideoTransformer3DModel
 
@@ -88,8 +88,7 @@ class LTXPipelineFastTests(PipelineTesterMixin, FirstBlockCacheTesterMixin, unit
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/ltx/test_ltx_condition.py

@@ -17,7 +17,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLLTXVideo,
@@ -92,8 +92,7 @@ class LTXConditionPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/ltx/test_ltx_image2video.py

@@ -17,7 +17,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLLTXVideo,
@@ -91,8 +91,7 @@ class LTXImageToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/mochi/test_mochi.py

@@ -18,7 +18,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLMochi, FlowMatchEulerDiscreteScheduler, MochiPipeline, MochiTransformer3DModel
 
@@ -89,8 +89,7 @@ class MochiPipelineFastTests(
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/pag/test_pag_hunyuan_dit.py

@@ -19,7 +19,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, BertModel, T5EncoderModel
+from transformers import AutoTokenizer, BertModel, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -67,9 +67,7 @@ class HunyuanDiTPAGPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         scheduler = DDPMScheduler()
         text_encoder = BertModel.from_pretrained("hf-internal-testing/tiny-random-BertModel")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-BertModel")
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         components = {

tests/pipelines/pag/test_pag_pixart_sigma.py

@@ -19,7 +19,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 import diffusers
 from diffusers import (
@@ -80,8 +80,7 @@ class PixArtSigmaPAGPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         vae = AutoencoderKL()
 
         scheduler = DDIMScheduler()
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 

tests/pipelines/pag/test_pag_sd3.py

@@ -3,14 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -80,9 +73,7 @@ class StableDiffusion3PAGPipelineFastTests(unittest.TestCase, PipelineTesterMixi
         torch.manual_seed(0)
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/pag/test_pag_sd3_img2img.py

@@ -5,14 +5,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -91,9 +84,7 @@ class StableDiffusion3PAGImg2ImgPipelineFastTests(unittest.TestCase, PipelineTes
         torch.manual_seed(0)
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/pixart_alpha/test_pixart.py

@@ -19,7 +19,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -77,10 +77,7 @@ class PixArtAlphaPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         vae = AutoencoderKL()
 
         scheduler = DDIMScheduler()
-
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 

tests/pipelines/pixart_sigma/test_pixart.py

@@ -19,7 +19,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -83,10 +83,7 @@ class PixArtSigmaPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         vae = AutoencoderKL()
 
         scheduler = DDIMScheduler()
-
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 

tests/pipelines/qwenimage/test_qwenimage.py

@@ -160,7 +160,7 @@ class QwenImagePipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         self.assertEqual(generated_image.shape, (3, 32, 32))
 
         # fmt: off
-        expected_slice = torch.tensor([0.5646, 0.6369, 0.6019, 0.5640, 0.5830, 0.5520, 0.5717, 0.6315, 0.4167, 0.3563, 0.5640, 0.4849, 0.4961, 0.5237, 0.4084, 0.5014])
+        expected_slice = torch.tensor([0.56331, 0.63677, 0.6015, 0.56369, 0.58166, 0.55277, 0.57176, 0.63261, 0.41466, 0.35561, 0.56229, 0.48334, 0.49714, 0.52622, 0.40872, 0.50208])
         # fmt: on
 
         generated_slice = generated_image.flatten()

tests/pipelines/qwenimage/test_qwenimage_edit.py

@@ -163,7 +163,7 @@ class QwenImageEditPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         self.assertEqual(generated_image.shape, (3, 32, 32))
 
         # fmt: off
-        expected_slice = torch.tensor([0.5640, 0.6350, 0.6003, 0.5606, 0.5801, 0.5502, 0.5757, 0.6388, 0.4174, 0.3590, 0.5647, 0.4891, 0.4975, 0.5256, 0.4088, 0.4991])
+        expected_slice = torch.tensor([[0.5637, 0.6341, 0.6001, 0.5620, 0.5794, 0.5498, 0.5757, 0.6389, 0.4174, 0.3597, 0.5649, 0.4894, 0.4969, 0.5255, 0.4083, 0.4986]])
         # fmt: on
 
         generated_slice = generated_image.flatten()

tests/pipelines/qwenimage/test_qwenimage_edit_plus.py

@@ -164,7 +164,7 @@ class QwenImageEditPlusPipelineFastTests(PipelineTesterMixin, unittest.TestCase)
         self.assertEqual(generated_image.shape, (3, 32, 32))
 
         # fmt: off
-        expected_slice = torch.tensor([0.5640, 0.6339, 0.5997, 0.5607, 0.5799, 0.5496, 0.5760, 0.6393, 0.4172, 0.3595, 0.5655, 0.4896, 0.4971, 0.5255, 0.4088, 0.4987])
+        expected_slice = torch.tensor([[0.5637, 0.6341, 0.6001, 0.5620, 0.5794, 0.5498, 0.5757, 0.6389, 0.4174, 0.3597, 0.5649, 0.4894, 0.4969, 0.5255, 0.4083, 0.4986]])
         # fmt: on
 
         generated_slice = generated_image.flatten()

tests/pipelines/skyreels_v2/test_skyreels_v2.py

@@ -16,7 +16,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLWan,
@@ -68,8 +68,7 @@ class SkyReelsV2PipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=8.0, use_flow_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/skyreels_v2/test_skyreels_v2_df.py

@@ -16,7 +16,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLWan,
@@ -68,8 +68,7 @@ class SkyReelsV2DiffusionForcingPipelineFastTests(PipelineTesterMixin, unittest.
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=8.0, use_flow_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/skyreels_v2/test_skyreels_v2_df_image_to_video.py

@@ -18,7 +18,6 @@ import numpy as np
 import torch
 from PIL import Image
 from transformers import (
-    AutoConfig,
     AutoTokenizer,
     T5EncoderModel,
 )
@@ -69,8 +68,7 @@ class SkyReelsV2DiffusionForcingImageToVideoPipelineFastTests(PipelineTesterMixi
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=5.0, use_flow_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
@@ -161,8 +159,7 @@ class SkyReelsV2DiffusionForcingImageToVideoPipelineFastTests(SkyReelsV2Diffusio
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=5.0, use_flow_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/skyreels_v2/test_skyreels_v2_df_video_to_video.py

@@ -18,7 +18,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLWan,
@@ -70,8 +70,7 @@ class SkyReelsV2DiffusionForcingVideoToVideoPipelineFastTests(PipelineTesterMixi
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=5.0, use_flow_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/skyreels_v2/test_skyreels_v2_image_to_video.py

@@ -18,7 +18,6 @@ import numpy as np
 import torch
 from PIL import Image
 from transformers import (
-    AutoConfig,
     AutoTokenizer,
     CLIPImageProcessor,
     CLIPVisionConfig,
@@ -72,8 +71,7 @@ class SkyReelsV2ImageToVideoPipelineFastTests(PipelineTesterMixin, unittest.Test
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=5.0, use_flow_sigmas=True)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/stable_audio/test_stable_audio.py

@@ -19,7 +19,10 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, T5EncoderModel, T5Tokenizer
+from transformers import (
+    T5EncoderModel,
+    T5Tokenizer,
+)
 
 from diffusers import (
     AutoencoderOobleck,
@@ -108,8 +111,7 @@ class StableAudioPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         )
         torch.manual_seed(0)
         t5_repo_id = "hf-internal-testing/tiny-random-T5ForConditionalGeneration"
-        config = AutoConfig.from_pretrained(t5_repo_id)
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained(t5_repo_id)
         tokenizer = T5Tokenizer.from_pretrained(t5_repo_id, truncation=True, model_max_length=25)
 
         torch.manual_seed(0)

tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3.py

@@ -3,14 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, SD3Transformer2DModel, StableDiffusion3Pipeline
 
@@ -79,9 +72,7 @@ class StableDiffusion3PipelineFastTests(unittest.TestCase, PipelineTesterMixin):
         torch.manual_seed(0)
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3_img2img.py

@@ -4,14 +4,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -80,9 +73,7 @@ class StableDiffusion3Img2ImgPipelineFastTests(PipelineLatentTesterMixin, unitte
         torch.manual_seed(0)
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/stable_diffusion_3/test_pipeline_stable_diffusion_3_inpaint.py

@@ -3,14 +3,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import (
-    AutoConfig,
-    AutoTokenizer,
-    CLIPTextConfig,
-    CLIPTextModelWithProjection,
-    CLIPTokenizer,
-    T5EncoderModel,
-)
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModelWithProjection, CLIPTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKL,
@@ -80,9 +73,7 @@ class StableDiffusion3InpaintPipelineFastTests(PipelineLatentTesterMixin, unitte
         torch.manual_seed(0)
         text_encoder_2 = CLIPTextModelWithProjection(clip_text_encoder_config)
 
-        torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_3 = T5EncoderModel(config)
+        text_encoder_3 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")

tests/pipelines/visualcloze/test_pipeline_visualcloze_combined.py

@@ -5,7 +5,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 import diffusers
 from diffusers import AutoencoderKL, FlowMatchEulerDiscreteScheduler, FluxTransformer2DModel, VisualClozePipeline
@@ -77,8 +77,7 @@ class VisualClozePipelineFastTests(unittest.TestCase, PipelineTesterMixin):
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/visualcloze/test_pipeline_visualcloze_generation.py

@@ -5,7 +5,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, CLIPTextConfig, CLIPTextModel, CLIPTokenizer, T5EncoderModel
 
 import diffusers
 from diffusers import (
@@ -79,8 +79,7 @@ class VisualClozeGenerationPipelineFastTests(unittest.TestCase, PipelineTesterMi
         text_encoder = CLIPTextModel(clip_text_encoder_config)
 
         torch.manual_seed(0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder_2 = T5EncoderModel(config)
+        text_encoder_2 = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         tokenizer = CLIPTokenizer.from_pretrained("hf-internal-testing/tiny-random-clip")
         tokenizer_2 = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")

tests/pipelines/wan/test_wan.py

@@ -18,7 +18,7 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLWan, FlowMatchEulerDiscreteScheduler, WanPipeline, WanTransformer3DModel
 
@@ -68,8 +68,7 @@ class WanPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         torch.manual_seed(0)
         # TODO: impl FlowDPMSolverMultistepScheduler
         scheduler = FlowMatchEulerDiscreteScheduler(shift=7.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/wan/test_wan_22.py

@@ -17,11 +17,14 @@ import unittest
 
 import numpy as np
 import torch
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLWan, UniPCMultistepScheduler, WanPipeline, WanTransformer3DModel
 
-from ...testing_utils import enable_full_determinism, torch_device
+from ...testing_utils import (
+    enable_full_determinism,
+    torch_device,
+)
 from ..pipeline_params import TEXT_TO_IMAGE_BATCH_PARAMS, TEXT_TO_IMAGE_IMAGE_PARAMS, TEXT_TO_IMAGE_PARAMS
 from ..test_pipelines_common import PipelineTesterMixin
 
@@ -60,8 +63,7 @@ class Wan22PipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(prediction_type="flow_prediction", use_flow_sigmas=True, flow_shift=3.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
@@ -233,8 +235,7 @@ class Wan225BPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(prediction_type="flow_prediction", use_flow_sigmas=True, flow_shift=3.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/wan/test_wan_22_image_to_video.py

@@ -18,7 +18,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLWan, UniPCMultistepScheduler, WanImageToVideoPipeline, WanTransformer3DModel
 
@@ -64,8 +64,7 @@ class Wan22ImageToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestCase)
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(prediction_type="flow_prediction", use_flow_sigmas=True, flow_shift=3.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
@@ -249,8 +248,7 @@ class Wan225BImageToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestCas
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(prediction_type="flow_prediction", use_flow_sigmas=True, flow_shift=3.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/wan/test_wan_animate.py

@@ -19,7 +19,6 @@ import numpy as np
 import torch
 from PIL import Image
 from transformers import (
-    AutoConfig,
     AutoTokenizer,
     CLIPImageProcessor,
     CLIPVisionConfig,
@@ -79,8 +78,7 @@ class WanAnimatePipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler(shift=7.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/wan/test_wan_image_to_video.py

@@ -19,7 +19,6 @@ import numpy as np
 import torch
 from PIL import Image
 from transformers import (
-    AutoConfig,
     AutoTokenizer,
     CLIPImageProcessor,
     CLIPVisionConfig,
@@ -69,8 +68,7 @@ class WanImageToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         torch.manual_seed(0)
         # TODO: impl FlowDPMSolverMultistepScheduler
         scheduler = FlowMatchEulerDiscreteScheduler(shift=7.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)
@@ -241,8 +239,7 @@ class WanFLFToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
         torch.manual_seed(0)
         # TODO: impl FlowDPMSolverMultistepScheduler
         scheduler = FlowMatchEulerDiscreteScheduler(shift=7.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/wan/test_wan_vace.py

@@ -18,7 +18,7 @@ import unittest
 import numpy as np
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import (
     AutoencoderKLWan,
@@ -67,8 +67,7 @@ class WanVACEPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = FlowMatchEulerDiscreteScheduler(shift=7.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)

tests/pipelines/wan/test_wan_video_to_video.py

@@ -16,7 +16,7 @@ import unittest
 
 import torch
 from PIL import Image
-from transformers import AutoConfig, AutoTokenizer, T5EncoderModel
+from transformers import AutoTokenizer, T5EncoderModel
 
 from diffusers import AutoencoderKLWan, UniPCMultistepScheduler, WanTransformer3DModel, WanVideoToVideoPipeline
 
@@ -62,8 +62,7 @@ class WanVideoToVideoPipelineFastTests(PipelineTesterMixin, unittest.TestCase):
 
         torch.manual_seed(0)
         scheduler = UniPCMultistepScheduler(flow_shift=3.0)
-        config = AutoConfig.from_pretrained("hf-internal-testing/tiny-random-t5")
-        text_encoder = T5EncoderModel(config)
+        text_encoder = T5EncoderModel.from_pretrained("hf-internal-testing/tiny-random-t5")
         tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/tiny-random-t5")
 
         torch.manual_seed(0)