[VLMs] support passing embeds along with pixels (#38467)

* VLMs can work with embeds now

* update more models

* fix tests

* fix copies

* fixup

* fix

* style

* unskip tests

* fix copies

* fix tests

* style

* omni modality models

* qwen models had extra indentation

* fix some other tests

* fix copies

* fix test last time

* unrelated changes revert

* we can't rely only on embeds

* delete file

* de-flake mistral3

* fix qwen models

* fix style

* fix tests

* fix copies

* deflake the test

* modular reverted by fixes, fix again

* flaky test, overwritten

* fix copies

* style

tests/models/gemma3/test_modeling_gemma3.py

@@ -131,10 +131,6 @@ class Gemma3ModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.TestCase
     def test_generate_from_inputs_embeds_with_static_cache(self):
         pass
 
-    @unittest.skip("Gemma3 has HybridCache and doesn't support StaticCache. Though it could, it shouldn't support.")
-    def test_generate_continue_from_inputs_embeds(self):
-        pass
-
     @unittest.skip("Gemma3 has HybridCache which auto-compiles. Compile and FA2 don't work together.")
     def test_eager_matches_fa2_generate(self):
         pass