add IP-Adapter support for SD 1.5

Official repo: https://github.com/tencent-ailab/IP-Adapter

scripts/conversion/convert_diffusers_ip_adapter.py

@@ -0,0 +1,111 @@
+from pathlib import Path
+from typing import Any
+import argparse
+
+import torch
+
+from refiners.foundationals.latent_diffusion import SD1UNet
+from refiners.foundationals.latent_diffusion.stable_diffusion_1 import SD1IPAdapter
+from refiners.fluxion.utils import save_to_safetensors
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(description="Converts a IP-Adapter diffusers model to refiners.")
+    parser.add_argument(
+        "--from",
+        type=str,
+        dest="source_path",
+        default="ip-adapter_sd15.bin",
+        help="Path to the source model. (default: 'ip-adapter_sd15.bin').",
+    )
+    parser.add_argument(
+        "--to",
+        type=str,
+        dest="output_path",
+        default="ip-adapter_sd15.safetensors",
+        help="Path to save the converted model. (default: 'ip-adapter_sd15.safetensors').",
+    )
+    parser.add_argument("--verbose", action="store_true", dest="verbose")
+    parser.add_argument("--half", action="store_true", dest="half")
+    args = parser.parse_args()
+    if args.output_path is None:
+        args.output_path = f"{Path(args.source_path).stem}.safetensors"
+
+    weights: dict[str, Any] = torch.load(f=args.source_path, map_location="cpu")  # type: ignore
+    assert isinstance(weights, dict)
+    assert sorted(weights.keys()) == ["image_proj", "ip_adapter"]
+
+    unet = SD1UNet(in_channels=4)
+
+    ip_adapter = SD1IPAdapter(target=unet)
+
+    # Manual conversion to avoid any runtime dependency on IP-Adapter[1] custom classes
+    # [1]: https://github.com/tencent-ailab/IP-Adapter
+
+    state_dict: dict[str, torch.Tensor] = {}
+
+    image_proj_weights = weights["image_proj"]
+    image_proj_state_dict: dict[str, torch.Tensor] = {
+        "Linear.weight": image_proj_weights["proj.weight"],
+        "Linear.bias": image_proj_weights["proj.bias"],
+        "LayerNorm.weight": image_proj_weights["norm.weight"],
+        "LayerNorm.bias": image_proj_weights["norm.bias"],
+    }
+    ip_adapter.image_proj.load_state_dict(state_dict=image_proj_state_dict)
+
+    for k, v in image_proj_state_dict.items():
+        state_dict[f"image_proj.{k}"] = v
+
+    ip_adapter_weights: dict[str, torch.Tensor] = weights["ip_adapter"]
+    assert len(ip_adapter.sub_adapters) == len(ip_adapter_weights.keys()) // 2
+
+    # Running:
+    #
+    #     from diffusers import UNet2DConditionModel
+    #     unet = UNet2DConditionModel.from_pretrained("runwayml/stable-diffusion-v1-5", subfolder="unet")
+    #     for k in unet.attn_processors.keys():
+    #         print(k)
+    #
+    # Gives:
+    #
+    #     down_blocks.0.attentions.0.transformer_blocks.0.attn1.processor
+    #     down_blocks.0.attentions.0.transformer_blocks.0.attn2.processor
+    #     ...
+    #     down_blocks.2.attentions.1.transformer_blocks.0.attn2.processor
+    #     up_blocks.1.attentions.0.transformer_blocks.0.attn1.processor
+    #     up_blocks.1.attentions.0.transformer_blocks.0.attn2.processor
+    #     ...
+    #     up_blocks.3.attentions.2.transformer_blocks.0.attn2.processor
+    #     mid_block.attentions.0.transformer_blocks.0.attn1.processor
+    #     mid_block.attentions.0.transformer_blocks.0.attn2.processor
+    #
+    # With attn1=self-attention and attn2=cross-attention, and middle block in last position. So in terms of increasing
+    # indices:
+    #
+    #     DownBlocks  -> [1, 3, 5, 7, 9, 11]
+    #     MiddleBlock -> [31]
+    #     UpBlocks    -> [13, 15, 17, 19, 21, 23, 25, 27, 29]
+    cross_attn_mapping: list[int] = [1, 3, 5, 7, 9, 11, 31, 13, 15, 17, 19, 21, 23, 25, 27, 29]
+
+    for i, cross_attn in enumerate(ip_adapter.sub_adapters):
+        cross_attn_index = cross_attn_mapping[i]
+        k_ip = f"{cross_attn_index}.to_k_ip.weight"
+        v_ip = f"{cross_attn_index}.to_v_ip.weight"
+
+        # Ignore Wq, Wk, Wv and Proj (hence strict=False): at runtime, they will be part of the UNet original weights
+        cross_attn_state_dict: dict[str, Any] = {
+            cross_attn.get_parameter_name("wk_prime"): ip_adapter_weights[k_ip],
+            cross_attn.get_parameter_name("wv_prime"): ip_adapter_weights[v_ip],
+        }
+        cross_attn.load_state_dict(state_dict=cross_attn_state_dict, strict=False)
+
+        for k, v in cross_attn_state_dict.items():
+            state_dict[f"ip_adapter.{i:03d}.{k}"] = v
+
+    if args.half:
+        state_dict = {key: value.half() for key, value in state_dict.items()}
+    save_to_safetensors(path=args.output_path, tensors=state_dict)
+
+
+if __name__ == "__main__":
+    main()

src/refiners/foundationals/latent_diffusion/__init__.py

@@ -10,6 +10,7 @@ from refiners.foundationals.latent_diffusion.stable_diffusion_1 import (
     StableDiffusion_1_Inpainting,
     SD1UNet,
     SD1ControlnetAdapter,
+    SD1IPAdapter,
 )
 from refiners.foundationals.latent_diffusion.stable_diffusion_xl import (
     SDXLUNet,
@@ -22,6 +23,7 @@ __all__ = [
     "StableDiffusion_1_Inpainting",
     "SD1UNet",
     "SD1ControlnetAdapter",
+    "SD1IPAdapter",
     "SDXLUNet",
     "DoubleTextEncoder",
     "DPMSolver",

src/refiners/foundationals/latent_diffusion/stable_diffusion_1/__init__.py

@@ -4,10 +4,12 @@ from refiners.foundationals.latent_diffusion.stable_diffusion_1.model import (
     StableDiffusion_1_Inpainting,
 )
 from refiners.foundationals.latent_diffusion.stable_diffusion_1.controlnet import SD1ControlnetAdapter
+from refiners.foundationals.latent_diffusion.stable_diffusion_1.image_prompt import SD1IPAdapter
 
 __all__ = [
     "StableDiffusion_1",
     "StableDiffusion_1_Inpainting",
     "SD1UNet",
     "SD1ControlnetAdapter",
+    "SD1IPAdapter",
 ]

src/refiners/foundationals/latent_diffusion/stable_diffusion_1/image_prompt.py

@@ -0,0 +1,348 @@
+from enum import IntEnum
+from pathlib import Path
+from functools import partial
+
+from torch import Tensor, as_tensor, cat, zeros_like, device as Device, dtype as DType
+from PIL import Image
+
+from refiners.fluxion.adapters.adapter import Adapter
+from refiners.foundationals.clip.image_encoder import CLIPImageEncoderH
+from refiners.foundationals.latent_diffusion.stable_diffusion_1.unet import SD1UNet
+from refiners.fluxion.layers.attentions import ScaledDotProductAttention
+from refiners.fluxion.utils import image_to_tensor, load_from_safetensors
+import refiners.fluxion.layers as fl
+
+
+class ImageProjection(fl.Chain):
+    structural_attrs = ["clip_image_embedding_dim", "clip_text_embedding_dim", "sequence_length"]
+
+    def __init__(
+        self,
+        clip_image_embedding_dim: int = 1024,
+        clip_text_embedding_dim: int = 768,
+        sequence_length: int = 4,
+        device: Device | str | None = None,
+        dtype: DType | None = None,
+    ) -> None:
+        self.clip_image_embedding_dim = clip_image_embedding_dim
+        self.clip_text_embedding_dim = clip_text_embedding_dim
+        self.sequence_length = sequence_length
+        super().__init__(
+            fl.Linear(
+                in_features=clip_image_embedding_dim,
+                out_features=clip_text_embedding_dim * sequence_length,
+                device=device,
+                dtype=dtype,
+            ),
+            fl.Reshape(sequence_length, clip_text_embedding_dim),
+            fl.LayerNorm(normalized_shape=clip_text_embedding_dim, device=device, dtype=dtype),
+        )
+
+
+class _CrossAttnIndex(IntEnum):
+    TXT_CROSS_ATTN = 0  # text cross-attention
+    IMG_CROSS_ATTN = 1  # image cross-attention
+
+
+# Fluxion's Attention layer drop-in replacement implementing Decoupled Cross-Attention
+class IPAttention(fl.Chain):
+    structural_attrs = [
+        "embedding_dim",
+        "text_sequence_length",
+        "image_sequence_length",
+        "scale",
+        "num_heads",
+        "heads_dim",
+        "key_embedding_dim",
+        "value_embedding_dim",
+        "inner_dim",
+        "use_bias",
+        "is_causal",
+    ]
+
+    def __init__(
+        self,
+        embedding_dim: int,
+        text_sequence_length: int = 77,
+        image_sequence_length: int = 4,
+        scale: float = 1.0,
+        num_heads: int = 1,
+        key_embedding_dim: int | None = None,
+        value_embedding_dim: int | None = None,
+        inner_dim: int | None = None,
+        use_bias: bool = True,
+        is_causal: bool | None = None,
+        device: Device | str | None = None,
+        dtype: DType | None = None,
+    ) -> None:
+        assert (
+            embedding_dim % num_heads == 0
+        ), f"embedding_dim {embedding_dim} must be divisible by num_heads {num_heads}"
+        self.embedding_dim = embedding_dim
+        self.text_sequence_length = text_sequence_length
+        self.image_sequence_length = image_sequence_length
+        self.scale = scale
+        self.num_heads = num_heads
+        self.heads_dim = embedding_dim // num_heads
+        self.key_embedding_dim = key_embedding_dim or embedding_dim
+        self.value_embedding_dim = value_embedding_dim or embedding_dim
+        self.inner_dim = inner_dim or embedding_dim
+        self.use_bias = use_bias
+        self.is_causal = is_causal
+        super().__init__(
+            fl.Distribute(
+                # Note: the same query is used for image cross-attention as for text cross-attention
+                fl.Linear(
+                    in_features=self.embedding_dim,
+                    out_features=self.inner_dim,
+                    bias=self.use_bias,
+                    device=device,
+                    dtype=dtype,
+                ),  # Wq
+                fl.Parallel(
+                    fl.Chain(
+                        fl.Slicing(dim=1, start=0, length=text_sequence_length),
+                        fl.Linear(
+                            in_features=self.key_embedding_dim,
+                            out_features=self.inner_dim,
+                            bias=self.use_bias,
+                            device=device,
+                            dtype=dtype,
+                        ),  # Wk
+                    ),
+                    fl.Chain(
+                        fl.Slicing(dim=1, start=text_sequence_length, length=image_sequence_length),
+                        fl.Linear(
+                            in_features=self.key_embedding_dim,
+                            out_features=self.inner_dim,
+                            bias=self.use_bias,
+                            device=device,
+                            dtype=dtype,
+                        ),  # Wk'
+                    ),
+                ),
+                fl.Parallel(
+                    fl.Chain(
+                        fl.Slicing(dim=1, start=0, length=text_sequence_length),
+                        fl.Linear(
+                            in_features=self.key_embedding_dim,
+                            out_features=self.inner_dim,
+                            bias=self.use_bias,
+                            device=device,
+                            dtype=dtype,
+                        ),  # Wv
+                    ),
+                    fl.Chain(
+                        fl.Slicing(dim=1, start=text_sequence_length, length=image_sequence_length),
+                        fl.Linear(
+                            in_features=self.key_embedding_dim,
+                            out_features=self.inner_dim,
+                            bias=self.use_bias,
+                            device=device,
+                            dtype=dtype,
+                        ),  # Wv'
+                    ),
+                ),
+            ),
+            fl.Sum(
+                fl.Chain(
+                    fl.Lambda(func=partial(self.select_qkv, index=_CrossAttnIndex.TXT_CROSS_ATTN)),
+                    ScaledDotProductAttention(num_heads=num_heads, is_causal=is_causal),
+                ),
+                fl.Chain(
+                    fl.Lambda(func=partial(self.select_qkv, index=_CrossAttnIndex.IMG_CROSS_ATTN)),
+                    ScaledDotProductAttention(num_heads=num_heads, is_causal=is_causal),
+                    fl.Lambda(func=self.scale_outputs),
+                ),
+            ),
+            fl.Linear(
+                in_features=self.inner_dim,
+                out_features=self.embedding_dim,
+                bias=True,
+                device=device,
+                dtype=dtype,
+            ),
+        )
+
+    def select_qkv(
+        self, query: Tensor, keys: tuple[Tensor, Tensor], values: tuple[Tensor, Tensor], index: _CrossAttnIndex
+    ) -> tuple[Tensor, Tensor, Tensor]:
+        return (query, keys[index.value], values[index.value])
+
+    def scale_outputs(self, x: Tensor) -> Tensor:
+        return x * self.scale
+
+
+class CrossAttentionAdapter(fl.Chain, Adapter[fl.Attention]):
+    structural_attrs = ["text_sequence_length", "image_sequence_length", "scale"]
+
+    def __init__(
+        self,
+        target: fl.Attention,
+        text_sequence_length: int = 77,
+        image_sequence_length: int = 4,
+        scale: float = 1.0,
+    ) -> None:
+        self.text_sequence_length = text_sequence_length
+        self.image_sequence_length = image_sequence_length
+        self.scale = scale
+        with self.setup_adapter(target):
+            super().__init__(
+                IPAttention(
+                    embedding_dim=target.embedding_dim,
+                    text_sequence_length=text_sequence_length,
+                    image_sequence_length=image_sequence_length,
+                    scale=scale,
+                    num_heads=target.num_heads,
+                    key_embedding_dim=target.key_embedding_dim,
+                    value_embedding_dim=target.value_embedding_dim,
+                    inner_dim=target.inner_dim,
+                    use_bias=target.use_bias,
+                    is_causal=target.is_causal,
+                    device=target.device,
+                    dtype=target.dtype,
+                )
+            )
+
+    def get_parameter_name(self, matrix: str, bias: bool = False) -> str:
+        match matrix:
+            case "wq":
+                index = 0
+            case "wk":
+                index = 1
+            case "wk_prime":
+                index = 2
+            case "wv":
+                index = 3
+            case "wv_prime":
+                index = 4
+            case "proj":
+                index = 5
+            case _:
+                raise ValueError(f"Unexpected matrix name {matrix}")
+
+        linear = list(self.IPAttention.layers(fl.Linear))[index]
+        param = getattr(linear, "bias" if bias else "weight")
+        name = next((n for n, p in self.named_parameters() if id(p) == id(param)), None)
+        assert name is not None
+        return name
+
+
+class SD1IPAdapter(fl.Chain, Adapter[SD1UNet]):
+    def __init__(
+        self,
+        target: SD1UNet,
+        clip_image_encoder: CLIPImageEncoderH | None = None,
+        scale: float = 1.0,
+        weights: dict[str, Tensor] | None = None,
+    ) -> None:
+        with self.setup_adapter(target):
+            super().__init__(target)
+
+        self.clip_image_encoder = clip_image_encoder or CLIPImageEncoderH(device=target.device, dtype=target.dtype)
+        self.image_proj = ImageProjection(device=target.device, dtype=target.dtype)
+
+        self.sub_adapters = [
+            CrossAttentionAdapter(target=cross_attn, scale=scale)
+            for cross_attn in filter(lambda attn: type(attn) != fl.SelfAttention, target.layers(fl.Attention))
+        ]
+
+        if weights is not None:
+            image_proj_state_dict: dict[str, Tensor] = {
+                k.removeprefix("image_proj."): v for k, v in weights.items() if k.startswith("image_proj.")
+            }
+            self.image_proj.load_state_dict(image_proj_state_dict)
+
+            for i, cross_attn in enumerate(self.sub_adapters):
+                cross_attn_state_dict: dict[str, Tensor] = {}
+                for k, v in weights.items():
+                    prefix = f"ip_adapter.{i:03d}."
+                    if not k.startswith(prefix):
+                        continue
+                    cross_attn_state_dict[k.removeprefix(prefix)] = v
+
+                # Retrieve original (frozen) cross-attention weights
+                # Note: this assumes the target UNet has already loaded weights
+                cross_attn_linears = list(cross_attn.target.layers(fl.Linear))
+                assert len(cross_attn_linears) == 4  # Wq, Wk, Wv and Proj
+
+                cross_attn_state_dict[cross_attn.get_parameter_name("wq")] = cross_attn_linears[0].weight
+                cross_attn_state_dict[cross_attn.get_parameter_name("wk")] = cross_attn_linears[1].weight
+                cross_attn_state_dict[cross_attn.get_parameter_name("wv")] = cross_attn_linears[2].weight
+                cross_attn_state_dict[cross_attn.get_parameter_name("proj")] = cross_attn_linears[3].weight
+                cross_attn_state_dict[cross_attn.get_parameter_name("proj", bias=True)] = cross_attn_linears[3].bias
+
+                cross_attn.load_state_dict(state_dict=cross_attn_state_dict)
+
+    @classmethod
+    def from_safetensors(
+        cls,
+        target: SD1UNet,
+        checkpoint_path: Path | str,
+        clip_image_encoder: CLIPImageEncoderH | None = None,
+        scale: float = 1.0,
+    ):
+        weights = load_from_safetensors(checkpoint_path, device=target.device if target.device is not None else "cpu")
+
+        return cls(
+            target=target,
+            clip_image_encoder=clip_image_encoder,
+            scale=scale,
+            weights=weights,
+        )
+
+    def inject(self: "SD1IPAdapter", parent: fl.Chain | None = None) -> "SD1IPAdapter":
+        for adapter in self.sub_adapters:
+            adapter.inject()
+        return super().inject(parent)
+
+    def eject(self) -> None:
+        for adapter in self.sub_adapters:
+            adapter.eject()
+        super().eject()
+
+    # These should be concatenated to the CLIP text embedding before setting the UNet context
+    def compute_clip_image_embedding(self, image_prompt: Tensor | None) -> Tensor:
+        clip_embedding = self.clip_image_encoder(image_prompt)
+        conditional_embedding = self.image_proj(clip_embedding)
+        negative_embedding = self.image_proj(zeros_like(clip_embedding))
+        return cat((negative_embedding, conditional_embedding))
+
+    def preprocess_image(
+        self,
+        image: Image.Image,
+        size: tuple[int, int] = (224, 224),
+        mean: list[float] | None = None,
+        std: list[float] | None = None,
+    ) -> Tensor:
+        # Default mean and std are parameters from https://github.com/openai/CLIP
+        return self._normalize(
+            image_to_tensor(image.resize(size), device=self.target.device, dtype=self.target.dtype),
+            mean=[0.48145466, 0.4578275, 0.40821073] if mean is None else mean,
+            std=[0.26862954, 0.26130258, 0.27577711] if std is None else std,
+        )
+
+    # Adapted from https://github.com/pytorch/vision/blob/main/torchvision/transforms/_functional_tensor.py
+    @staticmethod
+    def _normalize(tensor: Tensor, mean: list[float], std: list[float], inplace: bool = False) -> Tensor:
+        assert tensor.is_floating_point()
+        assert tensor.ndim >= 3
+
+        if not inplace:
+            tensor = tensor.clone()
+
+        dtype = tensor.dtype
+
+        mean_tensor = as_tensor(mean, dtype=tensor.dtype, device=tensor.device)
+        std_tensor = as_tensor(std, dtype=tensor.dtype, device=tensor.device)
+
+        if (std_tensor == 0).any():
+            raise ValueError(f"std evaluated to zero after conversion to {dtype}, leading to division by zero.")
+
+        if mean_tensor.ndim == 1:
+            mean_tensor = mean_tensor.view(-1, 1, 1)
+
+        if std_tensor.ndim == 1:
+            std_tensor = std_tensor.view(-1, 1, 1)
+
+        return tensor.sub_(mean_tensor).div_(std_tensor)

tests/e2e/test_diffusion.py

@@ -13,6 +13,7 @@ from refiners.foundationals.latent_diffusion import (
     StableDiffusion_1_Inpainting,
     SD1UNet,
     SD1ControlnetAdapter,
+    SD1IPAdapter,
 )
 from refiners.foundationals.latent_diffusion.lora import SD1LoraAdapter
 from refiners.foundationals.latent_diffusion.schedulers import DDIM
@@ -42,6 +43,11 @@ def kitchen_dog_mask(ref_path: Path) -> Image.Image:
     return Image.open(ref_path / "kitchen_dog_mask.png").convert("RGB")
 
 
+@pytest.fixture(scope="module")
+def woman_image(ref_path: Path) -> Image.Image:
+    return Image.open(ref_path / "woman.png").convert("RGB")
+
+
 @pytest.fixture
 def expected_image_std_random_init(ref_path: Path) -> Image.Image:
     return Image.open(ref_path / "expected_std_random_init.png").convert("RGB")
@@ -62,6 +68,11 @@ def expected_image_controlnet_stack(ref_path: Path) -> Image.Image:
     return Image.open(ref_path / "expected_controlnet_stack.png").convert("RGB")
 
 
+@pytest.fixture
+def expected_image_ip_adapter_woman(ref_path: Path) -> Image.Image:
+    return Image.open(ref_path / "expected_image_ip_adapter_woman.png").convert("RGB")
+
+
 @pytest.fixture(scope="module", params=["canny", "depth", "lineart", "normals", "sam"])
 def controlnet_data(
     ref_path: Path, test_weights_path: Path, request: pytest.FixtureRequest
@@ -182,6 +193,33 @@ def unet_weights_inpainting(test_weights_path: Path) -> Path:
     return unet_weights_inpainting
 
 
+@pytest.fixture(scope="module")
+def lda_ft_mse_weights(test_weights_path: Path) -> Path:
+    lda_weights = test_weights_path / "lda_ft_mse.safetensors"
+    if not lda_weights.is_file():
+        warn(f"could not find weights at {lda_weights}, skipping")
+        pytest.skip(allow_module_level=True)
+    return lda_weights
+
+
+@pytest.fixture(scope="module")
+def ip_adapter_weights(test_weights_path: Path) -> Path:
+    ip_adapter_weights = test_weights_path / "ip-adapter_sd15.safetensors"
+    if not ip_adapter_weights.is_file():
+        warn(f"could not find weights at {ip_adapter_weights}, skipping")
+        pytest.skip(allow_module_level=True)
+    return ip_adapter_weights
+
+
+@pytest.fixture(scope="module")
+def image_encoder_weights(test_weights_path: Path) -> Path:
+    image_encoder_weights = test_weights_path / "CLIPImageEncoderH.safetensors"
+    if not image_encoder_weights.is_file():
+        warn(f"could not find weights at {image_encoder_weights}, skipping")
+        pytest.skip(allow_module_level=True)
+    return image_encoder_weights
+
+
 @pytest.fixture
 def sd15_std(
     text_encoder_weights: Path, lda_weights: Path, unet_weights_std: Path, test_device: torch.device
@@ -270,6 +308,24 @@ def sd15_ddim(
     return sd15
 
 
+@pytest.fixture
+def sd15_ddim_lda_ft_mse(
+    text_encoder_weights: Path, lda_ft_mse_weights: Path, unet_weights_std: Path, test_device: torch.device
+) -> StableDiffusion_1:
+    if test_device.type == "cpu":
+        warn("not running on CPU, skipping")
+        pytest.skip()
+
+    ddim_scheduler = DDIM(num_inference_steps=20)
+    sd15 = StableDiffusion_1(scheduler=ddim_scheduler, device=test_device)
+
+    sd15.clip_text_encoder.load_state_dict(load_from_safetensors(text_encoder_weights))
+    sd15.lda.load_state_dict(load_from_safetensors(lda_ft_mse_weights))
+    sd15.unet.load_state_dict(load_from_safetensors(unet_weights_std))
+
+    return sd15
+
+
 @torch.no_grad()
 def test_diffusion_std_random_init(
     sd15_std: StableDiffusion_1, expected_image_std_random_init: Image.Image, test_device: torch.device
@@ -844,3 +900,60 @@ def test_diffusion_textual_inversion_random_init(
         predicted_image = sd15.lda.decode_latents(x)
 
     ensure_similar_images(predicted_image, expected_image_textual_inversion_random_init, min_psnr=35, min_ssim=0.98)
+
+
+@torch.no_grad()
+def test_diffusion_ip_adapter(
+    sd15_ddim_lda_ft_mse: StableDiffusion_1,
+    ip_adapter_weights: Path,
+    image_encoder_weights: Path,
+    woman_image: Image.Image,
+    expected_image_ip_adapter_woman: Image.Image,
+    test_device: torch.device,
+):
+    sd15 = sd15_ddim_lda_ft_mse.to(dtype=torch.float16)
+    n_steps = 50
+
+    # See tencent-ailab/IP-Adapter best practices section:
+    #
+    #     If you only use the image prompt, you can set the scale=1.0 and text_prompt="" (or some generic text
+    #     prompts, e.g. "best quality", you can also use any negative text prompt).
+    #
+    # The prompts below are the ones used by default by IPAdapter's generate method if none are specified
+    prompt = "best quality, high quality"
+    negative_prompt = "monochrome, lowres, bad anatomy, worst quality, low quality"
+
+    ip_adapter = SD1IPAdapter.from_safetensors(target=sd15.unet, checkpoint_path=ip_adapter_weights)
+    ip_adapter.clip_image_encoder.load_from_safetensors(image_encoder_weights)
+    ip_adapter.inject()
+
+    with torch.no_grad():
+        clip_text_embedding = sd15.compute_clip_text_embedding(text=prompt, negative_text=negative_prompt)
+        clip_image_embedding = ip_adapter.compute_clip_image_embedding(ip_adapter.preprocess_image(woman_image))
+
+        negative_text_embedding, conditional_text_embedding = clip_text_embedding.chunk(2)
+        negative_image_embedding, conditional_image_embedding = clip_image_embedding.chunk(2)
+
+        clip_text_embedding = torch.cat(
+            (
+                torch.cat([negative_text_embedding, negative_image_embedding], dim=1),
+                torch.cat([conditional_text_embedding, conditional_image_embedding], dim=1),
+            )
+        )
+
+    sd15.set_num_inference_steps(n_steps)
+
+    manual_seed(2)
+    x = torch.randn(1, 4, 64, 64, device=test_device, dtype=torch.float16)
+
+    with torch.no_grad():
+        for step in sd15.steps:
+            x = sd15(
+                x,
+                step=step,
+                clip_text_embedding=clip_text_embedding,
+                condition_scale=7.5,
+            )
+        predicted_image = sd15.lda.decode_latents(x)
+
+    ensure_similar_images(predicted_image, expected_image_ip_adapter_woman)

tests/e2e/test_diffusion_ref/README.md

@@ -35,7 +35,7 @@ output.images[0].save("std_random_init_expected.png")
 Special cases:
 
 - `expected_refonly.png` has been generated [with Stable Diffusion web UI](https://github.com/AUTOMATIC1111/stable-diffusion-webui).
-- `expected_inpainting_refonly.png` has been generated with refiners itself (and inspected so that it looks reasonable).
+- `expected_inpainting_refonly.png`, `expected_image_ip_adapter_woman.png` have been generated with refiners itself (and inspected so that it looks reasonable).
 
 ## Other images
 
@@ -59,6 +59,8 @@ Special cases:
     - `inpainting-scene.png`: cropped-to-square-and-resized version of https://unsplash.com/photos/RCz6eSVPGYU by @jannerboy62
     - `inpainting-target.png`: computed with `convert <(convert -size 512x512 xc:white png:-) kitchen_dog.png <(convert inpainting-mask.png -negate png:-) -compose Over -composite inpainting-target.png`
 
+- `woman.png` [comes from tencent-ailab/IP-Adapter](https://github.com/tencent-ailab/IP-Adapter/blob/8b96670cc5c8ef00278b42c0c7b62fe8a74510b9/assets/images/woman.png).
+
 ## VAE without randomness
 
 ```diff