[Attention] R3 Attention Transform (#485)

Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>

Author: kylesayrs

src/compressed_tensors/transform/factory/base.py

@@ -18,6 +18,14 @@
 import torch
 import torch.nn.utils.parametrize as P
 import tqdm
+from compressed_tensors.modeling.attention import (
+    initialize_hooked_attention,
+    register_query_hook,
+)
+from compressed_tensors.modeling.kvcache import (
+    initialize_hooked_kv_cache,
+    register_key_hook,
+)
 from compressed_tensors.registry.registry import RegistryMixin, T
 from compressed_tensors.transform import (
     TransformArgs,
@@ -36,6 +44,7 @@
 from compressed_tensors.utils.internal import InternalModule
 from torch import Tensor
 from torch.nn import Module, Parameter
+from transformers import PreTrainedModel
 
 
 __all__ = ["TransformFactory", "TransformBase"]
@@ -97,12 +106,13 @@ def apply_to_model(self, model: Module, use_tqdm=True):
 
         desc = f"Applying {self.name} transforms"
         for module, arg in tqdm.tqdm(modules_args, desc=desc, disable=(not use_tqdm)):
-            self._apply_to_module(module, arg)
+            self._apply_to_module(model, module, arg)
 
-    def _apply_to_module(self, module: Module, args: TransformArgs):
+    def _apply_to_module(self, model: Module, module: Module, args: TransformArgs):
         """
         Create transforms and apply them to the module
 
+        :param model: model which module belongs to
         :param module: target module to apply transforms to
         :param args: defines how the transform will be applied to the target module
         """
@@ -156,7 +166,28 @@ def output_hook(_, _input, output):
 
             module.register_forward_hook(output_hook)
 
-        # other locations such as q_attn and k_attn have not been implemented
+        # register query hook to attention
+        elif args.location == TransformLocation.Q_ATTN:
+            if not isinstance(model, PreTrainedModel):
+                raise ValueError(f"Cannot hook attention of model: {model}")
+
+            def query_hook(_, query_states):
+                return transform(query_states)
+
+            initialize_hooked_attention(model, module)
+            register_query_hook(module, query_hook)
+
+        # register key hook to kvcache
+        elif args.location == TransformLocation.K_CACHE:
+            if not isinstance(model, PreTrainedModel):
+                raise ValueError(f"Cannot hook attention of model: {model}")
+
+            def key_hook(_, key_states):
+                return transform(key_states)
+
+            initialize_hooked_kv_cache(model, module)
+            register_key_hook(module, key_hook)
+
         else:
             raise NotImplementedError()
 

src/compressed_tensors/transform/factory/hadamard.py

@@ -51,7 +51,6 @@ def create_transform(self, module: Module, args: TransformArgs):
         :param module: parent module that transform will be applied to
         :param args: defines how the transform will be applied to the module
         """
-        assert hasattr(module, "weight")
         size = get_transform_size(module, args.location, self.scheme.head_dim)
         exec_device = get_execution_device(module)
         device = get_offloaded_device(module)

src/compressed_tensors/transform/factory/matrix_multiply.py

@@ -50,7 +50,6 @@ def create_transform(self, module: Module, args: TransformArgs):
         :param module: parent module that transform will be applied to
         :param args: defines how the transform will be applied to the module
         """
-        assert hasattr(module, "weight")
         size = get_transform_size(module, args.location, self.scheme.head_dim)
         device = get_offloaded_device(module)
         precision = self.scheme.precision if args.is_online() else torch.float64

src/compressed_tensors/transform/transform_args.py

@@ -45,6 +45,16 @@ class TransformLocation(str, Enum):
     K_CACHE = "k_cache"
     Q_ATTN = "q_attn"
 
+    def is_online(self) -> bool:
+        """
+        Returns True if the transform location is online
+        (applied at runtime), False otherwise
+        """
+        return self not in (
+            TransformLocation.WEIGHT_INPUT,
+            TransformLocation.WEIGHT_OUTPUT,
+        )
+
 
 class TransformArgs(BaseModel, use_enum_values=True):
     """
@@ -70,9 +80,6 @@ def wrap_singleton(cls, value):
         return value
 
     def is_online(self) -> bool:
-        return self.location not in (
-            TransformLocation.WEIGHT_INPUT,
-            TransformLocation.WEIGHT_OUTPUT,
-        )
+        return TransformLocation(self.location).is_online()
 
     model_config = ConfigDict(extra="forbid")

src/compressed_tensors/transform/utils/matrix.py

@@ -34,6 +34,8 @@ def get_transform_size(
     :param head_dim: size of head when transform is applied to mha
     :return: size of matrix
     """
+    size = None
+
     if isinstance(module, torch.nn.Linear):
         if location in (TransformLocation.INPUT, TransformLocation.WEIGHT_INPUT):
             size = module.in_features
@@ -44,11 +46,13 @@ def get_transform_size(
             size = module.num_embeddings
         else:
             size = module.embedding_dim
-    else:
-        raise NotImplementedError(f"Transforms on {type(module)} are not supported")
+    elif head_dim is None:
+        raise NotImplementedError(
+            f"Transforms on {type(module)} are not supported without head_dim"
+        )
 
     if head_dim is not None:
-        if size % head_dim != 0:
+        if size is not None and size % head_dim != 0:
             raise ValueError(
                 f"{head_dim} must divide {size} for {type(module)} at {location}"
             )
@@ -105,38 +109,26 @@ def apply_transform_weight(
 
     assert transform_weight.shape[0] == transform_weight.shape[1]
 
-    if module_type == torch.nn.Linear:
-        if location == TransformLocation.INPUT:
-            return _multihead_matmul(value, transform_weight)
+    if TransformLocation(location).is_online():
+        return _multihead_matmul(value, transform_weight)
 
-        elif location == TransformLocation.WEIGHT_INPUT:
+    if module_type == torch.nn.Linear:
+        if location == TransformLocation.WEIGHT_INPUT:
             # equivalent to (transform_weight @ value.T).T
             return _multihead_matmul(value, transform_weight.T)
 
         elif location == TransformLocation.WEIGHT_OUTPUT:
             # equivalent to (value.T @ transform_weight).T
             return _multihead_matmul(transform_weight.T, value)
 
-        elif location == TransformLocation.OUTPUT:
-            return _multihead_matmul(value, transform_weight)
-
     # similar derivation to torch.nn.Linear, but `y = (x W)`
     elif module_type == torch.nn.Embedding:
-        if location == TransformLocation.INPUT:
-            return _multihead_matmul(value, transform_weight)
-
-        elif location == TransformLocation.WEIGHT_INPUT:
-            return _multihead_matmul(
-                transform_weight,
-                value,
-            )
+        if location == TransformLocation.WEIGHT_INPUT:
+            return _multihead_matmul(transform_weight, value)
 
         elif location == TransformLocation.WEIGHT_OUTPUT:
             return _multihead_matmul(value, transform_weight)
 
-        elif location == TransformLocation.OUTPUT:
-            return _multihead_matmul(value, transform_weight)
-
     raise NotImplementedError(
         f"Applying transforms to {module_type} {location} is not supported"
     )

tests/test_transform/conftest.py

@@ -62,14 +62,19 @@ def __init__(
             num_attention_heads * self.head_dim, hidden_size, bias=False
         )
 
-    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
+    def forward(
+        self, hidden_states: torch.Tensor, past_key_values=None
+    ) -> torch.Tensor:
         batch_size, seq_len, hidden_size = hidden_states.shape
         hidden_shape = (batch_size, seq_len, -1, self.head_dim)
 
         query_states = self.q_proj(hidden_states).view(hidden_shape).transpose(1, 2)
         key_states = self.k_proj(hidden_states).view(hidden_shape).transpose(1, 2)
         value_states = self.v_proj(hidden_states).view(hidden_shape).transpose(1, 2)
 
+        if past_key_values is not None:
+            past_key_values.update(key_states, value_states, 0, {})
+
         key_states = self.repeat_kv(key_states, self.num_key_value_groups)
         value_states = self.repeat_kv(value_states, self.num_key_value_groups)
 
@@ -97,6 +102,21 @@ def repeat_kv(self, hidden_states: torch.Tensor, n_rep: int) -> torch.Tensor:
         return hidden_states.reshape(batch, num_key_value_heads * n_rep, slen, head_dim)
 
 
+class MockAttentionModel(PreTrainedModel):
+    config_class = PretrainedConfig
+
+    def __init__(self, hidden_size, num_attention_heads, num_key_value_heads):
+        super().__init__(PretrainedConfig())
+        self.self_attn = MockAttention(
+            hidden_size=hidden_size,
+            num_attention_heads=num_attention_heads,
+            num_key_value_heads=num_key_value_heads,
+        )
+
+    def forward(self, x):
+        return self.self_attn(x)
+
+
 @pytest.fixture(scope="function")
 def model_apply():
     model = TransformableModel(2, 4, 8, 16, 32, 64)

tests/test_transform/factory/test_correctness.py

@@ -22,7 +22,7 @@
     apply_transform_config,
 )
 from compressed_tensors.utils import offloaded_dispatch
-from tests.test_transform.conftest import MockAttention
+from tests.test_transform.conftest import MockAttention, MockAttentionModel
 from tests.testing_utils import requires_accelerate, requires_gpu
 
 
@@ -147,7 +147,7 @@ def test_correctness_attention_heads(type, randomize, head_dim, input_batch_size
 
     config = TransformConfig(
         config_groups={
-            "": TransformScheme(
+            "R2": TransformScheme(
                 type=type,
                 randomize=randomize,
                 head_dim=head_dim,
@@ -166,6 +166,42 @@ def test_correctness_attention_heads(type, randomize, head_dim, input_batch_size
     assert torch.allclose(true_output, output, atol=1e-5, rtol=0.0)
 
 
+@pytest.mark.parametrize("type", ("hadamard", "random-hadamard"))
+@pytest.mark.parametrize("randomize", (True, False))
+@pytest.mark.parametrize("head_dim", (4, 8))
+@pytest.mark.parametrize("input_batch_size", (1, 5, 17))
+def test_correctness_query_key_locations(type, randomize, head_dim, input_batch_size):
+    hidden_size = 64
+    num_attention_heads = 8
+
+    model = MockAttentionModel(
+        hidden_size=hidden_size,
+        num_attention_heads=num_attention_heads,
+        num_key_value_heads=head_dim,
+    )
+
+    input = torch.rand(input_batch_size, 5, hidden_size)
+    true_output = model(input)
+
+    config = TransformConfig(
+        config_groups={
+            "R3": TransformScheme(
+                type=type,
+                randomize=randomize,
+                head_dim=head_dim,
+                apply=[
+                    TransformArgs(targets="self_attn", location="q_attn"),
+                    TransformArgs(targets="self_attn", location="k_cache"),
+                ],
+            )
+        }
+    )
+    apply_transform_config(model, config)
+
+    output = model(input)
+    assert torch.allclose(true_output, output, atol=1e-5, rtol=0.0)
+
+
 @requires_gpu
 @pytest.mark.parametrize("cuda_default", (True, False))
 def test_random_matrix_device_handling(cuda_default):