🔴 VLM: compile compatibility (#35724)

* llavas

* add mroe models

* fix `compile_forward` test for all models

* fix copies

* make style

* also doesn't support cache class

* fix some tests

* not copied from

* ci green?

* fix tests

* fix copies

* fix tests

* check with `numel` and remove `item`

* fix copies

* fix copies

* Update src/transformers/models/cohere2/modeling_cohere2.py

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* opt remove cross attn

* gemma2

* fixup

* fixup

* fix newly added test

* maybe fixed?

* green please?

---------

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

tests/models/opt/test_modeling_opt.py

@@ -81,7 +81,7 @@ class OPTModelTester:
         hidden_act="gelu",
         hidden_dropout_prob=0.1,
         attention_probs_dropout_prob=0.1,
-        max_position_embeddings=20,
+        max_position_embeddings=50,
         eos_token_id=2,
         pad_token_id=1,
         bos_token_id=0,
@@ -89,7 +89,6 @@ class OPTModelTester:
         num_labels=3,
         word_embed_proj_dim=16,
         type_sequence_label_size=2,
-        attn_implementation="eager",
     ):
         self.parent = parent
         self.batch_size = batch_size
@@ -113,7 +112,6 @@ class OPTModelTester:
         self.type_sequence_label_size = type_sequence_label_size
         self.word_embed_proj_dim = word_embed_proj_dim
         self.is_encoder_decoder = False
-        self.attn_implementation = attn_implementation
 
     def prepare_config_and_inputs(self):
         input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size).clamp(
@@ -143,7 +141,6 @@ class OPTModelTester:
             embed_dim=self.embed_dim,
             is_encoder_decoder=False,
             word_embed_proj_dim=self.word_embed_proj_dim,
-            attn_implementation=self.attn_implementation,
         )
 
     def get_pipeline_config(self):