Merge branch 'main' into model-doc_vit_mae

Author: mreraser

tests/models/phimoe/test_modeling_phimoe.py

@@ -102,6 +102,7 @@ class PhimoeModelTest(CausalLMModelTest, unittest.TestCase):
 
     test_headmasking = False
     test_pruning = False
+    test_all_params_have_gradient = False
     model_tester_class = PhimoeModelTester
     pipeline_model_mapping = (
         {

tests/models/qwen2_moe/test_modeling_qwen2_moe.py

@@ -83,6 +83,7 @@ class Qwen2MoeModelTest(CausalLMModelTest, unittest.TestCase):
 
     test_headmasking = False
     test_pruning = False
+    test_all_params_have_gradient = False
     model_tester_class = Qwen2MoeModelTester
 
     # TODO (ydshieh): Check this. See https://app.circleci.com/pipelines/github/huggingface/transformers/79245/workflows/9490ef58-79c2-410d-8f51-e3495156cf9c/jobs/1012146

tests/models/qwen3_moe/test_modeling_qwen3_moe.py

@@ -82,6 +82,7 @@ class Qwen3MoeModelTest(CausalLMModelTest, unittest.TestCase):
 
     test_headmasking = False
     test_pruning = False
+    test_all_params_have_gradient = False
     model_tester_class = Qwen3MoeModelTester
 
     # TODO (ydshieh): Check this. See https://app.circleci.com/pipelines/github/huggingface/transformers/79245/workflows/9490ef58-79c2-410d-8f51-e3495156cf9c/jobs/1012146