Stable Video Diffusion

sgm/modules/attention.py

@@ -1,3 +1,4 @@
+import logging
 import math
 from inspect import isfunction
 from typing import Any, Optional
@@ -7,6 +8,9 @@ import torch.nn.functional as F
 from einops import rearrange, repeat
 from packaging import version
 from torch import nn
+from torch.utils.checkpoint import checkpoint
+
+logpy = logging.getLogger(__name__)
 
 if version.parse(torch.__version__) >= version.parse("2.0.0"):
     SDP_IS_AVAILABLE = True
@@ -36,9 +40,10 @@ else:
     SDP_IS_AVAILABLE = False
     sdp_kernel = nullcontext
     BACKEND_MAP = {}
-    print(
-        f"No SDP backend available, likely because you are running in pytorch versions < 2.0. In fact, "
-        f"you are using PyTorch {torch.__version__}. You might want to consider upgrading."
+    logpy.warn(
+        f"No SDP backend available, likely because you are running in pytorch "
+        f"versions < 2.0. In fact, you are using PyTorch {torch.__version__}. "
+        f"You might want to consider upgrading."
     )
 
 try:
@@ -48,9 +53,9 @@ try:
     XFORMERS_IS_AVAILABLE = True
 except:
     XFORMERS_IS_AVAILABLE = False
-    print("no module 'xformers'. Processing without...")
+    logpy.warn("no module 'xformers'. Processing without...")
 
-from .diffusionmodules.util import checkpoint
+# from .diffusionmodules.util import mixed_checkpoint as checkpoint
 
 
 def exists(val):
@@ -146,6 +151,62 @@ class LinearAttention(nn.Module):
         return self.to_out(out)
 
 
+class SelfAttention(nn.Module):
+    ATTENTION_MODES = ("xformers", "torch", "math")
+
+    def __init__(
+        self,
+        dim: int,
+        num_heads: int = 8,
+        qkv_bias: bool = False,
+        qk_scale: Optional[float] = None,
+        attn_drop: float = 0.0,
+        proj_drop: float = 0.0,
+        attn_mode: str = "xformers",
+    ):
+        super().__init__()
+        self.num_heads = num_heads
+        head_dim = dim // num_heads
+        self.scale = qk_scale or head_dim**-0.5
+
+        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
+        self.attn_drop = nn.Dropout(attn_drop)
+        self.proj = nn.Linear(dim, dim)
+        self.proj_drop = nn.Dropout(proj_drop)
+        assert attn_mode in self.ATTENTION_MODES
+        self.attn_mode = attn_mode
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        B, L, C = x.shape
+
+        qkv = self.qkv(x)
+        if self.attn_mode == "torch":
+            qkv = rearrange(
+                qkv, "B L (K H D) -> K B H L D", K=3, H=self.num_heads
+            ).float()
+            q, k, v = qkv[0], qkv[1], qkv[2]  # B H L D
+            x = torch.nn.functional.scaled_dot_product_attention(q, k, v)
+            x = rearrange(x, "B H L D -> B L (H D)")
+        elif self.attn_mode == "xformers":
+            qkv = rearrange(qkv, "B L (K H D) -> K B L H D", K=3, H=self.num_heads)
+            q, k, v = qkv[0], qkv[1], qkv[2]  # B L H D
+            x = xformers.ops.memory_efficient_attention(q, k, v)
+            x = rearrange(x, "B L H D -> B L (H D)", H=self.num_heads)
+        elif self.attn_mode == "math":
+            qkv = rearrange(qkv, "B L (K H D) -> K B H L D", K=3, H=self.num_heads)
+            q, k, v = qkv[0], qkv[1], qkv[2]  # B H L D
+            attn = (q @ k.transpose(-2, -1)) * self.scale
+            attn = attn.softmax(dim=-1)
+            attn = self.attn_drop(attn)
+            x = (attn @ v).transpose(1, 2).reshape(B, L, C)
+        else:
+            raise NotImplemented
+
+        x = self.proj(x)
+        x = self.proj_drop(x)
+        return x
+
+
 class SpatialSelfAttention(nn.Module):
     def __init__(self, in_channels):
         super().__init__()
@@ -289,9 +350,10 @@ class MemoryEfficientCrossAttention(nn.Module):
         self, query_dim, context_dim=None, heads=8, dim_head=64, dropout=0.0, **kwargs
     ):
         super().__init__()
-        print(
-            f"Setting up {self.__class__.__name__}. Query dim is {query_dim}, context_dim is {context_dim} and using "
-            f"{heads} heads with a dimension of {dim_head}."
+        logpy.debug(
+            f"Setting up {self.__class__.__name__}. Query dim is {query_dim}, "
+            f"context_dim is {context_dim} and using {heads} heads with a "
+            f"dimension of {dim_head}."
         )
         inner_dim = dim_head * heads
         context_dim = default(context_dim, query_dim)
@@ -352,9 +414,29 @@ class MemoryEfficientCrossAttention(nn.Module):
         )
 
         # actually compute the attention, what we cannot get enough of
-        out = xformers.ops.memory_efficient_attention(
-            q, k, v, attn_bias=None, op=self.attention_op
-        )
+        if version.parse(xformers.__version__) >= version.parse("0.0.21"):
+            # NOTE: workaround for
+            # https://github.com/facebookresearch/xformers/issues/845
+            max_bs = 32768
+            N = q.shape[0]
+            n_batches = math.ceil(N / max_bs)
+            out = list()
+            for i_batch in range(n_batches):
+                batch = slice(i_batch * max_bs, (i_batch + 1) * max_bs)
+                out.append(
+                    xformers.ops.memory_efficient_attention(
+                        q[batch],
+                        k[batch],
+                        v[batch],
+                        attn_bias=None,
+                        op=self.attention_op,
+                    )
+                )
+            out = torch.cat(out, 0)
+        else:
+            out = xformers.ops.memory_efficient_attention(
+                q, k, v, attn_bias=None, op=self.attention_op
+            )
 
         # TODO: Use this directly in the attention operation, as a bias
         if exists(mask):
@@ -393,21 +475,24 @@ class BasicTransformerBlock(nn.Module):
         super().__init__()
         assert attn_mode in self.ATTENTION_MODES
         if attn_mode != "softmax" and not XFORMERS_IS_AVAILABLE:
-            print(
-                f"Attention mode '{attn_mode}' is not available. Falling back to native attention. "
-                f"This is not a problem in Pytorch >= 2.0. FYI, you are running with PyTorch version {torch.__version__}"
+            logpy.warn(
+                f"Attention mode '{attn_mode}' is not available. Falling "
+                f"back to native attention. This is not a problem in "
+                f"Pytorch >= 2.0. FYI, you are running with PyTorch "
+                f"version {torch.__version__}."
             )
             attn_mode = "softmax"
         elif attn_mode == "softmax" and not SDP_IS_AVAILABLE:
-            print(
-                "We do not support vanilla attention anymore, as it is too expensive. Sorry."
+            logpy.warn(
+                "We do not support vanilla attention anymore, as it is too "
+                "expensive. Sorry."
             )
             if not XFORMERS_IS_AVAILABLE:
                 assert (
                     False
                 ), "Please install xformers via e.g. 'pip install xformers==0.0.16'"
             else:
-                print("Falling back to xformers efficient attention.")
+                logpy.info("Falling back to xformers efficient attention.")
                 attn_mode = "softmax-xformers"
         attn_cls = self.ATTENTION_MODES[attn_mode]
         if version.parse(torch.__version__) >= version.parse("2.0.0"):
@@ -437,7 +522,7 @@ class BasicTransformerBlock(nn.Module):
         self.norm3 = nn.LayerNorm(dim)
         self.checkpoint = checkpoint
         if self.checkpoint:
-            print(f"{self.__class__.__name__} is using checkpointing")
+            logpy.debug(f"{self.__class__.__name__} is using checkpointing")
 
     def forward(
         self, x, context=None, additional_tokens=None, n_times_crossframe_attn_in_self=0
@@ -456,9 +541,12 @@ class BasicTransformerBlock(nn.Module):
             )
 
         # return mixed_checkpoint(self._forward, kwargs, self.parameters(), self.checkpoint)
-        return checkpoint(
-            self._forward, (x, context), self.parameters(), self.checkpoint
-        )
+        if self.checkpoint:
+            # inputs = {"x": x, "context": context}
+            return checkpoint(self._forward, x, context)
+            # return checkpoint(self._forward, inputs, self.parameters(), self.checkpoint)
+        else:
+            return self._forward(**kwargs)
 
     def _forward(
         self, x, context=None, additional_tokens=None, n_times_crossframe_attn_in_self=0
@@ -518,9 +606,9 @@ class BasicTransformerSingleLayerBlock(nn.Module):
         self.checkpoint = checkpoint
 
     def forward(self, x, context=None):
-        return checkpoint(
-            self._forward, (x, context), self.parameters(), self.checkpoint
-        )
+        # inputs = {"x": x, "context": context}
+        # return checkpoint(self._forward, inputs, self.parameters(), self.checkpoint)
+        return checkpoint(self._forward, x, context)
 
     def _forward(self, x, context=None):
         x = self.attn1(self.norm1(x), context=context) + x
@@ -554,18 +642,20 @@ class SpatialTransformer(nn.Module):
         sdp_backend=None,
     ):
         super().__init__()
-        print(
-            f"constructing {self.__class__.__name__} of depth {depth} w/ {in_channels} channels and {n_heads} heads"
+        logpy.debug(
+            f"constructing {self.__class__.__name__} of depth {depth} w/ "
+            f"{in_channels} channels and {n_heads} heads."
         )
-        from omegaconf import ListConfig
 
-        if exists(context_dim) and not isinstance(context_dim, (list, ListConfig)):
+        if exists(context_dim) and not isinstance(context_dim, list):
             context_dim = [context_dim]
         if exists(context_dim) and isinstance(context_dim, list):
             if depth != len(context_dim):
-                print(
-                    f"WARNING: {self.__class__.__name__}: Found context dims {context_dim} of depth {len(context_dim)}, "
-                    f"which does not match the specified 'depth' of {depth}. Setting context_dim to {depth * [context_dim[0]]} now."
+                logpy.warn(
+                    f"{self.__class__.__name__}: Found context dims "
+                    f"{context_dim} of depth {len(context_dim)}, which does not "
+                    f"match the specified 'depth' of {depth}. Setting context_dim "
+                    f"to {depth * [context_dim[0]]} now."
                 )
                 # depth does not match context dims.
                 assert all(
@@ -631,3 +721,39 @@ class SpatialTransformer(nn.Module):
         if not self.use_linear:
             x = self.proj_out(x)
         return x + x_in
+
+
+class SimpleTransformer(nn.Module):
+    def __init__(
+        self,
+        dim: int,
+        depth: int,
+        heads: int,
+        dim_head: int,
+        context_dim: Optional[int] = None,
+        dropout: float = 0.0,
+        checkpoint: bool = True,
+    ):
+        super().__init__()
+        self.layers = nn.ModuleList([])
+        for _ in range(depth):
+            self.layers.append(
+                BasicTransformerBlock(
+                    dim,
+                    heads,
+                    dim_head,
+                    dropout=dropout,
+                    context_dim=context_dim,
+                    attn_mode="softmax-xformers",
+                    checkpoint=checkpoint,
+                )
+            )
+
+    def forward(
+        self,
+        x: torch.Tensor,
+        context: Optional[torch.Tensor] = None,
+    ) -> torch.Tensor:
+        for layer in self.layers:
+            x = layer(x, context)
+        return x