F.scaled_dot_product_attention support (#26572)

* add sdpa

* wip

* cleaning

* add ref

* yet more cleaning

* and more :)

* wip llama

* working llama

* add output_attentions=True support

* bigcode sdpa support

* fixes

* gpt-bigcode support, require torch>=2.1.1

* add falcon support

* fix conflicts falcon

* style

* fix attention_mask definition

* remove output_attentions from attnmaskconverter

* support whisper without removing any Copied from statement

* fix mbart default to eager renaming

* fix typo in falcon

* fix is_causal in SDPA

* check is_flash_attn_2_available in the models init as well in case the model is not initialized through from_pretrained

* add warnings when falling back on the manual implementation

* precise doc

* wip replace _flash_attn_enabled by config.attn_implementation

* fix typo

* add tests

* style

* add a copy.deepcopy on the config in from_pretrained, as we do not want to modify it inplace

* obey to config.attn_implementation if a config is passed in from_pretrained

* fix is_torch_sdpa_available when torch is not installed

* remove dead code

* Update src/transformers/modeling_attn_mask_utils.py

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

* Update src/transformers/modeling_attn_mask_utils.py

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

* Update src/transformers/modeling_attn_mask_utils.py

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

* Update src/transformers/modeling_attn_mask_utils.py

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

* Update src/transformers/modeling_attn_mask_utils.py

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

* Update src/transformers/models/bart/modeling_bart.py

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

* remove duplicate pretraining_tp code

* add dropout in llama

* precise comment on attn_mask

* add fmt: off for _unmask_unattended docstring

* precise num_masks comment

* nuke pretraining_tp in LlamaSDPAAttention following Arthur's suggestion

* cleanup modeling_utils

* backward compatibility

* fix style as requested

* style

* improve documentation

* test pass

* style

* add _unmask_unattended tests

* skip meaningless tests for idefics

* hard_check SDPA requirements when specifically requested

* standardize the use if XXX_ATTENTION_CLASSES

* fix SDPA bug with mem-efficient backend on CUDA when using fp32

* fix test

* rely on SDPA is_causal parameter to handle the causal mask in some cases

* fix FALCON_ATTENTION_CLASSES

* remove _flash_attn_2_enabled occurences

* fix test

* add OPT to the list of supported flash models

* improve test

* properly test on different SDPA backends, on different dtypes & properly handle separately the pad tokens in the test

* remove remaining _flash_attn_2_enabled occurence

* Update src/transformers/modeling_utils.py

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

* Update src/transformers/modeling_utils.py

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

* Update src/transformers/modeling_utils.py

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

* Update src/transformers/modeling_attn_mask_utils.py

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

* Update docs/source/en/perf_infer_gpu_one.md

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

* remove use_attn_implementation

* fix docstring & slight bug

* make attn_implementation internal (_attn_implementation)

* typos

* fix tests

* deprecate use_flash_attention_2=True

* fix test

* add back llama that was removed by mistake

* fix tests

* remove _flash_attn_2_enabled occurences bis

* add check & test that passed attn_implementation is valid

* fix falcon torchscript export

* fix device of mask in tests

* add tip about torch.jit.trace and move bt doc below sdpa

* fix parameterized.expand order

* move tests from test_modeling_attn_mask_utils to test_modeling_utils as a relevant test class is already there

* update sdpaattention class with the new cache

* Update src/transformers/configuration_utils.py

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

* Update src/transformers/models/bark/modeling_bark.py

* address review comments

* WIP torch.jit.trace fix. left: test both eager & sdpa

* add test for torch.jit.trace for both eager/sdpa

* fix falcon with torch==2.0 that needs to use sdpa

* fix doc

* hopefully last fix

* fix key_value_length that has no default now in mask converter

* is it flacky?

* fix speculative decoding bug

* tests do pass

* fix following #27907

---------

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

tests/models/idefics/test_modeling_idefics.py

@@ -16,11 +16,14 @@
 
 import unittest
 
+from parameterized import parameterized
+
 from transformers import BitsAndBytesConfig, IdeficsConfig, is_torch_available, is_vision_available
 from transformers.testing_utils import (
     TestCasePlus,
     require_bitsandbytes,
     require_torch,
+    require_torch_sdpa,
     require_vision,
     slow,
     torch_device,
@@ -309,6 +312,12 @@ class IdeficsModelTester:
     def prepare_pixel_values(self):
         return floats_tensor([self.batch_size, self.num_channels, self.image_size, self.image_size])
 
+    @require_torch_sdpa
+    @slow
+    @parameterized.expand([("float16",), ("bfloat16",), ("float32",)])
+    def test_eager_matches_sdpa_inference(self, torch_dtype: str):
+        self.skipTest("Idefics has a hard requirement on SDPA, skipping this test")
+
 
 @unittest.skipIf(not is_torch_greater_or_equal_than_2_0, reason="pytorch 2.0 or higher is required")
 @require_torch
@@ -557,6 +566,12 @@ class IdeficsModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase)
             model = IdeficsModel.from_pretrained(model_name)
             self.assertIsNotNone(model)
 
+    @require_torch_sdpa
+    @slow
+    @parameterized.expand([("float16",), ("bfloat16",), ("float32",)])
+    def test_eager_matches_sdpa_inference(self, torch_dtype: str):
+        self.skipTest("Idefics has a hard requirement on SDPA, skipping this test")
+
 
 @unittest.skipIf(not is_torch_greater_or_equal_than_2_0, reason="pytorch 2.0 or higher is required")
 @require_torch