🚨🚨🚨 Enforce single model initialization (#21431)

* Enforce single model initialization

* Add OneFormer example for problem 3

* Do it the Stas way

* Actually rename the uses...

* Rewrite test

* Try to change the test this way

* Fix all init slow/fast tests

* Break connection

* Fix more tests

* Fix test for initialization

* Remove custom test

* Quality

* Fix last failing tests

* The end?

tests/models/dpt/test_modeling_dpt.py

@@ -24,7 +24,7 @@ from transformers.models.auto import get_values
 from transformers.testing_utils import require_torch, require_vision, slow, torch_device
 
 from ...test_configuration_common import ConfigTester
-from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
+from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
 
 
 if is_torch_available():
@@ -242,6 +242,29 @@ class DPTModelTest(ModelTesterMixin, unittest.TestCase):
             loss = model(**inputs).loss
             loss.backward()
 
+    def test_initialization(self):
+        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
+
+        configs_no_init = _config_zero_init(config)
+        for model_class in self.all_model_classes:
+            model = model_class(config=configs_no_init)
+            # Skip the check for the backbone
+            backbone_params = []
+            for name, module in model.named_modules():
+                if module.__class__.__name__ == "DPTViTHybridEmbeddings":
+                    backbone_params = [f"{name}.{key}" for key in module.state_dict().keys()]
+                    break
+
+            for name, param in model.named_parameters():
+                if param.requires_grad:
+                    if name in backbone_params:
+                        continue
+                    self.assertIn(
+                        ((param.data.mean() * 1e9).round() / 1e9).item(),
+                        [0.0, 1.0],
+                        msg=f"Parameter {name} of model {model_class} seems not properly initialized",
+                    )
+
     @slow
     def test_model_from_pretrained(self):
         for model_name in DPT_PRETRAINED_MODEL_ARCHIVE_LIST[:1]:

tests/models/dpt/test_modeling_dpt_hybrid.py

@@ -24,7 +24,7 @@ from transformers.models.auto import get_values
 from transformers.testing_utils import require_torch, require_vision, slow, torch_device
 
 from ...test_configuration_common import ConfigTester
-from ...test_modeling_common import ModelTesterMixin, floats_tensor, ids_tensor
+from ...test_modeling_common import ModelTesterMixin, _config_zero_init, floats_tensor, ids_tensor
 
 
 if is_torch_available():
@@ -256,6 +256,29 @@ class DPTModelTest(ModelTesterMixin, unittest.TestCase):
             loss = model(**inputs).loss
             loss.backward()
 
+    def test_initialization(self):
+        config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common()
+
+        configs_no_init = _config_zero_init(config)
+        for model_class in self.all_model_classes:
+            model = model_class(config=configs_no_init)
+            # Skip the check for the backbone
+            backbone_params = []
+            for name, module in model.named_modules():
+                if module.__class__.__name__ == "DPTViTHybridEmbeddings":
+                    backbone_params = [f"{name}.{key}" for key in module.state_dict().keys()]
+                    break
+
+            for name, param in model.named_parameters():
+                if param.requires_grad:
+                    if name in backbone_params:
+                        continue
+                    self.assertIn(
+                        ((param.data.mean() * 1e9).round() / 1e9).item(),
+                        [0.0, 1.0],
+                        msg=f"Parameter {name} of model {model_class} seems not properly initialized",
+                    )
+
     @slow
     def test_model_from_pretrained(self):
         for model_name in DPT_PRETRAINED_MODEL_ARCHIVE_LIST[1:]: