[Core] fuse_qkv_projection() to Flux (#9185)

* start fusing flux.

* test

* finish fusion

* fix-copues

Author: sayakpaul

src/diffusers/models/attention_processor.py

@@ -1783,6 +1783,100 @@ def __call__(
             return hidden_states
 
 
+class FusedFluxAttnProcessor2_0:
+    """Attention processor used typically in processing the SD3-like self-attention projections."""
+
+    def __init__(self):
+        if not hasattr(F, "scaled_dot_product_attention"):
+            raise ImportError(
+                "FusedFluxAttnProcessor2_0 requires PyTorch 2.0, to use it, please upgrade PyTorch to 2.0."
+            )
+
+    def __call__(
+        self,
+        attn: Attention,
+        hidden_states: torch.FloatTensor,
+        encoder_hidden_states: torch.FloatTensor = None,
+        attention_mask: Optional[torch.FloatTensor] = None,
+        image_rotary_emb: Optional[torch.Tensor] = None,
+    ) -> torch.FloatTensor:
+        batch_size, _, _ = hidden_states.shape if encoder_hidden_states is None else encoder_hidden_states.shape
+
+        # `sample` projections.
+        qkv = attn.to_qkv(hidden_states)
+        split_size = qkv.shape[-1] // 3
+        query, key, value = torch.split(qkv, split_size, dim=-1)
+
+        inner_dim = key.shape[-1]
+        head_dim = inner_dim // attn.heads
+
+        query = query.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
+        key = key.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
+        value = value.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
+
+        if attn.norm_q is not None:
+            query = attn.norm_q(query)
+        if attn.norm_k is not None:
+            key = attn.norm_k(key)
+
+        # the attention in FluxSingleTransformerBlock does not use `encoder_hidden_states`
+        # `context` projections.
+        if encoder_hidden_states is not None:
+            encoder_qkv = attn.to_added_qkv(encoder_hidden_states)
+            split_size = encoder_qkv.shape[-1] // 3
+            (
+                encoder_hidden_states_query_proj,
+                encoder_hidden_states_key_proj,
+                encoder_hidden_states_value_proj,
+            ) = torch.split(encoder_qkv, split_size, dim=-1)
+
+            encoder_hidden_states_query_proj = encoder_hidden_states_query_proj.view(
+                batch_size, -1, attn.heads, head_dim
+            ).transpose(1, 2)
+            encoder_hidden_states_key_proj = encoder_hidden_states_key_proj.view(
+                batch_size, -1, attn.heads, head_dim
+            ).transpose(1, 2)
+            encoder_hidden_states_value_proj = encoder_hidden_states_value_proj.view(
+                batch_size, -1, attn.heads, head_dim
+            ).transpose(1, 2)
+
+            if attn.norm_added_q is not None:
+                encoder_hidden_states_query_proj = attn.norm_added_q(encoder_hidden_states_query_proj)
+            if attn.norm_added_k is not None:
+                encoder_hidden_states_key_proj = attn.norm_added_k(encoder_hidden_states_key_proj)
+
+            # attention
+            query = torch.cat([encoder_hidden_states_query_proj, query], dim=2)
+            key = torch.cat([encoder_hidden_states_key_proj, key], dim=2)
+            value = torch.cat([encoder_hidden_states_value_proj, value], dim=2)
+
+        if image_rotary_emb is not None:
+            from .embeddings import apply_rotary_emb
+
+            query = apply_rotary_emb(query, image_rotary_emb)
+            key = apply_rotary_emb(key, image_rotary_emb)
+
+        hidden_states = F.scaled_dot_product_attention(query, key, value, dropout_p=0.0, is_causal=False)
+        hidden_states = hidden_states.transpose(1, 2).reshape(batch_size, -1, attn.heads * head_dim)
+        hidden_states = hidden_states.to(query.dtype)
+
+        if encoder_hidden_states is not None:
+            encoder_hidden_states, hidden_states = (
+                hidden_states[:, : encoder_hidden_states.shape[1]],
+                hidden_states[:, encoder_hidden_states.shape[1] :],
+            )
+
+            # linear proj
+            hidden_states = attn.to_out[0](hidden_states)
+            # dropout
+            hidden_states = attn.to_out[1](hidden_states)
+            encoder_hidden_states = attn.to_add_out(encoder_hidden_states)
+
+            return hidden_states, encoder_hidden_states
+        else:
+            return hidden_states
+
+
 class CogVideoXAttnProcessor2_0:
     r"""
     Processor for implementing scaled dot-product attention for the CogVideoX model. It applies a rotary embedding on

src/diffusers/models/transformers/transformer_flux.py

@@ -23,7 +23,12 @@
 from ...configuration_utils import ConfigMixin, register_to_config
 from ...loaders import FromOriginalModelMixin, PeftAdapterMixin
 from ...models.attention import FeedForward
-from ...models.attention_processor import Attention, FluxAttnProcessor2_0
+from ...models.attention_processor import (
+    Attention,
+    AttentionProcessor,
+    FluxAttnProcessor2_0,
+    FusedFluxAttnProcessor2_0,
+)
 from ...models.modeling_utils import ModelMixin
 from ...models.normalization import AdaLayerNormContinuous, AdaLayerNormZero, AdaLayerNormZeroSingle
 from ...utils import USE_PEFT_BACKEND, is_torch_version, logging, scale_lora_layers, unscale_lora_layers
@@ -276,6 +281,106 @@ def __init__(
 
         self.gradient_checkpointing = False
 
+    @property
+    # Copied from diffusers.models.unets.unet_2d_condition.UNet2DConditionModel.attn_processors
+    def attn_processors(self) -> Dict[str, AttentionProcessor]:
+        r"""
+        Returns:
+            `dict` of attention processors: A dictionary containing all attention processors used in the model with
+            indexed by its weight name.
+        """
+        # set recursively
+        processors = {}
+
+        def fn_recursive_add_processors(name: str, module: torch.nn.Module, processors: Dict[str, AttentionProcessor]):
+            if hasattr(module, "get_processor"):
+                processors[f"{name}.processor"] = module.get_processor()
+
+            for sub_name, child in module.named_children():
+                fn_recursive_add_processors(f"{name}.{sub_name}", child, processors)
+
+            return processors
+
+        for name, module in self.named_children():
+            fn_recursive_add_processors(name, module, processors)
+
+        return processors
+
+    # Copied from diffusers.models.unets.unet_2d_condition.UNet2DConditionModel.set_attn_processor
+    def set_attn_processor(self, processor: Union[AttentionProcessor, Dict[str, AttentionProcessor]]):
+        r"""
+        Sets the attention processor to use to compute attention.
+
+        Parameters:
+            processor (`dict` of `AttentionProcessor` or only `AttentionProcessor`):
+                The instantiated processor class or a dictionary of processor classes that will be set as the processor
+                for **all** `Attention` layers.
+
+                If `processor` is a dict, the key needs to define the path to the corresponding cross attention
+                processor. This is strongly recommended when setting trainable attention processors.
+
+        """
+        count = len(self.attn_processors.keys())
+
+        if isinstance(processor, dict) and len(processor) != count:
+            raise ValueError(
+                f"A dict of processors was passed, but the number of processors {len(processor)} does not match the"
+                f" number of attention layers: {count}. Please make sure to pass {count} processor classes."
+            )
+
+        def fn_recursive_attn_processor(name: str, module: torch.nn.Module, processor):
+            if hasattr(module, "set_processor"):
+                if not isinstance(processor, dict):
+                    module.set_processor(processor)
+                else:
+                    module.set_processor(processor.pop(f"{name}.processor"))
+
+            for sub_name, child in module.named_children():
+                fn_recursive_attn_processor(f"{name}.{sub_name}", child, processor)
+
+        for name, module in self.named_children():
+            fn_recursive_attn_processor(name, module, processor)
+
+    # Copied from diffusers.models.unets.unet_2d_condition.UNet2DConditionModel.fuse_qkv_projections with FusedAttnProcessor2_0->FusedFluxAttnProcessor2_0
+    def fuse_qkv_projections(self):
+        """
+        Enables fused QKV projections. For self-attention modules, all projection matrices (i.e., query, key, value)
+        are fused. For cross-attention modules, key and value projection matrices are fused.
+
+        <Tip warning={true}>
+
+        This API is 🧪 experimental.
+
+        </Tip>
+        """
+        self.original_attn_processors = None
+
+        for _, attn_processor in self.attn_processors.items():
+            if "Added" in str(attn_processor.__class__.__name__):
+                raise ValueError("`fuse_qkv_projections()` is not supported for models having added KV projections.")
+
+        self.original_attn_processors = self.attn_processors
+
+        for module in self.modules():
+            if isinstance(module, Attention):
+                module.fuse_projections(fuse=True)
+
+        self.set_attn_processor(FusedFluxAttnProcessor2_0())
+
+    # Copied from diffusers.models.unets.unet_2d_condition.UNet2DConditionModel.unfuse_qkv_projections
+    def unfuse_qkv_projections(self):
+        """Disables the fused QKV projection if enabled.
+
+        <Tip warning={true}>
+
+        This API is 🧪 experimental.
+
+        </Tip>
+
+        """
+        if self.original_attn_processors is not None:
+            self.set_attn_processor(self.original_attn_processors)
+
     def _set_gradient_checkpointing(self, module, value=False):
         if hasattr(module, "gradient_checkpointing"):
             module.gradient_checkpointing = value

tests/pipelines/flux/test_pipeline_flux.py

@@ -13,10 +13,13 @@
     torch_device,
 )
 
-from ..test_pipelines_common import PipelineTesterMixin
+from ..test_pipelines_common import (
+    PipelineTesterMixin,
+    check_qkv_fusion_matches_attn_procs_length,
+    check_qkv_fusion_processors_exist,
+)
 
 
-@unittest.skipIf(torch_device == "mps", "Flux has a float64 operation which is not supported in MPS.")
 class FluxPipelineFastTests(unittest.TestCase, PipelineTesterMixin):
     pipeline_class = FluxPipeline
     params = frozenset(["prompt", "height", "width", "guidance_scale", "prompt_embeds", "pooled_prompt_embeds"])
@@ -143,6 +146,46 @@ def test_flux_prompt_embeds(self):
         max_diff = np.abs(output_with_prompt - output_with_embeds).max()
         assert max_diff < 1e-4
 
+    def test_fused_qkv_projections(self):
+        device = "cpu"  # ensure determinism for the device-dependent torch.Generator
+        components = self.get_dummy_components()
+        pipe = self.pipeline_class(**components)
+        pipe = pipe.to(device)
+        pipe.set_progress_bar_config(disable=None)
+
+        inputs = self.get_dummy_inputs(device)
+        image = pipe(**inputs).images
+        original_image_slice = image[0, -3:, -3:, -1]
+
+        # TODO (sayakpaul): will refactor this once `fuse_qkv_projections()` has been added
+        # to the pipeline level.
+        pipe.transformer.fuse_qkv_projections()
+        assert check_qkv_fusion_processors_exist(
+            pipe.transformer
+        ), "Something wrong with the fused attention processors. Expected all the attention processors to be fused."
+        assert check_qkv_fusion_matches_attn_procs_length(
+            pipe.transformer, pipe.transformer.original_attn_processors
+        ), "Something wrong with the attention processors concerning the fused QKV projections."
+
+        inputs = self.get_dummy_inputs(device)
+        image = pipe(**inputs).images
+        image_slice_fused = image[0, -3:, -3:, -1]
+
+        pipe.transformer.unfuse_qkv_projections()
+        inputs = self.get_dummy_inputs(device)
+        image = pipe(**inputs).images
+        image_slice_disabled = image[0, -3:, -3:, -1]
+
+        assert np.allclose(
+            original_image_slice, image_slice_fused, atol=1e-3, rtol=1e-3
+        ), "Fusion of QKV projections shouldn't affect the outputs."
+        assert np.allclose(
+            image_slice_fused, image_slice_disabled, atol=1e-3, rtol=1e-3
+        ), "Outputs, with QKV projection fusion enabled, shouldn't change when fused QKV projections are disabled."
+        assert np.allclose(
+            original_image_slice, image_slice_disabled, atol=1e-2, rtol=1e-2
+        ), "Original outputs should match when fused QKV projections are disabled."
+
 
 @slow
 @require_torch_gpu