Merge branch 'main' into improve-security-stylebot

.github/workflows/pr_style_bot.yml

@@ -64,18 +64,38 @@ jobs:
         run: |
           pip install .[quality]
 
-      - name: Download Makefile from main branch
+      - name: Download necessary files from main branch of Diffusers
         run: |
           curl -o main_Makefile https://raw.githubusercontent.com/huggingface/diffusers/main/Makefile
+          curl -o main_setup.py https://raw.githubusercontent.com/huggingface/diffusers/refs/heads/main/setup.py
+          curl -o main_check_doc_toc.py https://raw.githubusercontent.com/huggingface/diffusers/refs/heads/main/utils/check_doc_toc.py
         
-      - name: Compare Makefiles
+      - name: Compare the files and raise error if needed
         run: |
+          diff_failed=0
+
           if ! diff -q main_Makefile Makefile; then
             echo "Error: The Makefile has changed. Please ensure it matches the main branch."
+            diff_failed=1
+          fi
+
+          if ! diff -q main_setup.py setup.py; then
+            echo "Error: The setup.py has changed. Please ensure it matches the main branch."
+            diff_failed=1
+          fi
+
+          if ! diff -q main_check_doc_toc.py utils/check_doc_toc.py; then
+            echo "Error: The utils/check_doc_toc.py has changed. Please ensure it matches the main branch."
+            diff_failed=1
+          fi
+
+          if [ $diff_failed -eq 1 ]; then
+            echo "❌ Error happened as we detected changes in the files that should not be changed ❌"
             exit 1
           fi
-          echo "No changes in Makefile. Proceeding..."
-          rm -rf main_Makefile
+
+          echo "No changes in the files. Proceeding..."
+          rm -rf main_Makefile main_setup.py main_check_doc_toc.py
 
       - name: Run make style and make quality
         run: |

.github/workflows/pr_tests_gpu.yml

@@ -11,8 +11,6 @@ on:
       - "src/diffusers/loaders/lora_base.py"
       - "src/diffusers/loaders/lora_pipeline.py"
       - "src/diffusers/loaders/peft.py"
-      - "tests/pipelines/test_pipelines_common.py"
-      - "tests/models/test_modeling_common.py"
   workflow_dispatch:
 
 concurrency:

tests/models/test_modeling_common.py

@@ -1169,16 +1169,17 @@ class ModelTesterMixin:
         base_output = model(**inputs_dict)
 
         model_size = compute_module_sizes(model)[""]
-        max_size = int(self.model_split_percents[0] * model_size)
-        # Force disk offload by setting very small CPU memory
-        max_memory = {0: max_size, "cpu": int(0.1 * max_size)}
-
         with tempfile.TemporaryDirectory() as tmp_dir:
             model.cpu().save_pretrained(tmp_dir, safe_serialization=False)
+
             with self.assertRaises(ValueError):
+                max_size = int(self.model_split_percents[0] * model_size)
+                max_memory = {0: max_size, "cpu": max_size}
                 # This errors out because it's missing an offload folder
                 new_model = self.model_class.from_pretrained(tmp_dir, device_map="auto", max_memory=max_memory)
 
+            max_size = int(self.model_split_percents[0] * model_size)
+            max_memory = {0: max_size, "cpu": max_size}
             new_model = self.model_class.from_pretrained(
                 tmp_dir, device_map="auto", max_memory=max_memory, offload_folder=tmp_dir
             )

tests/models/transformers/test_models_transformer_omnigen.py

@@ -30,7 +30,6 @@ class OmniGenTransformerTests(ModelTesterMixin, unittest.TestCase):
     model_class = OmniGenTransformer2DModel
     main_input_name = "hidden_states"
     uses_custom_attn_processor = True
-    model_split_percents = [0.1, 0.1, 0.1]
 
     @property
     def dummy_input(self):
@@ -74,9 +73,9 @@ class OmniGenTransformerTests(ModelTesterMixin, unittest.TestCase):
             "num_attention_heads": 4,
             "num_key_value_heads": 4,
             "intermediate_size": 32,
-            "num_layers": 20,
+            "num_layers": 1,
             "pad_token_id": 0,
-            "vocab_size": 1000,
+            "vocab_size": 100,
             "in_channels": 4,
             "time_step_dim": 4,
             "rope_scaling": {"long_factor": list(range(1, 3)), "short_factor": list(range(1, 3))},

tests/models/transformers/test_models_transformer_sd3.py

@@ -33,7 +33,6 @@ enable_full_determinism()
 class SD3TransformerTests(ModelTesterMixin, unittest.TestCase):
     model_class = SD3Transformer2DModel
     main_input_name = "hidden_states"
-    model_split_percents = [0.8, 0.8, 0.9]
 
     @property
     def dummy_input(self):
@@ -68,7 +67,7 @@ class SD3TransformerTests(ModelTesterMixin, unittest.TestCase):
             "sample_size": 32,
             "patch_size": 1,
             "in_channels": 4,
-            "num_layers": 4,
+            "num_layers": 1,
             "attention_head_dim": 8,
             "num_attention_heads": 4,
             "caption_projection_dim": 32,
@@ -108,7 +107,6 @@ class SD3TransformerTests(ModelTesterMixin, unittest.TestCase):
 class SD35TransformerTests(ModelTesterMixin, unittest.TestCase):
     model_class = SD3Transformer2DModel
     main_input_name = "hidden_states"
-    model_split_percents = [0.8, 0.8, 0.9]
 
     @property
     def dummy_input(self):
@@ -143,7 +141,7 @@ class SD35TransformerTests(ModelTesterMixin, unittest.TestCase):
             "sample_size": 32,
             "patch_size": 1,
             "in_channels": 4,
-            "num_layers": 4,
+            "num_layers": 2,
             "attention_head_dim": 8,
             "num_attention_heads": 4,
             "caption_projection_dim": 32,