[Model] Gemma3n MM (#20495)

Signed-off-by: ShriKode <shrikode@gmail.com>
Signed-off-by: NickLucche <nlucches@redhat.com>
Signed-off-by: Roger Wang <hey@rogerw.me>
Co-authored-by: ShriKode <shrikode@gmail.com>
Co-authored-by: Roger Wang <hey@rogerw.me>

tests/models/multimodal/processing/test_common.py

@@ -271,6 +271,7 @@ def _test_processing_correctness_one(
     "microsoft/Florence-2-base",
     "adept/fuyu-8b",
     "google/gemma-3-4b-it",
+    "google/gemma-3n-E2B-it",
     "zai-org/glm-4v-9b",
     "zai-org/GLM-4.1V-9B-Thinking",
     "ibm-granite/granite-speech-3.3-2b",
@@ -315,7 +316,7 @@ def _test_processing_correctness_one(
     "fixie-ai/ultravox-v0_5-llama-3_2-1b",
     "openai/whisper-large-v3",
     "omni-research/Tarsier-7b",
-    "omni-research/Tarsier2-Recap-7b"
+    "omni-research/Tarsier2-Recap-7b",
 ])
 @pytest.mark.parametrize("hit_rate", [0.3, 0.5, 1.0])
 @pytest.mark.parametrize("num_batches", [32])
@@ -327,6 +328,8 @@ def test_processing_correctness(
     num_batches: int,
     simplify_rate: float,
 ):
+    if model_id == "google/gemma-3n-E2B-it":
+        pytest.skip("Skipping gemma-3n-E2B-it due to transformers #39911 bug.")
     _test_processing_correctness(
         model_id,
         hit_rate=hit_rate,