mlx-examples/segment_anything/segment_anything/image_encoder.py

from typing import Optional, Tuple, Type

import mlx.core as mx
import mlx.nn as nn

from .common import LayerNorm2d, MLPBlock


class ImageEncoderViT(nn.Module):
    def __init__(
        self,
        img_size: int = 1024,
        patch_size: int = 16,
        in_chans: int = 3,
        embed_dim: int = 768,
        depth: int = 12,
        num_heads: int = 12,
        mlp_ratio: float = 4.0,
        out_chans: int = 256,
        qkv_bias: bool = True,
        norm_layer: Type[nn.Module] = nn.LayerNorm,
        act_layer: Type[nn.Module] = nn.GELU,
        use_abs_pos: bool = True,
        use_rel_pos: bool = False,
        rel_pos_zero_init: bool = True,
        window_size: int = 0,
        global_attn_indexes: Tuple[int, ...] = (),
    ) -> None:
        """
        Args:
            img_size (int): Input image size.
            patch_size (int): Patch size.
            in_chans (int): Number of input image channels.
            embed_dim (int): Patch embedding dimension.
            depth (int): Depth of ViT.
            num_heads (int): Number of attention heads in each ViT block.
            mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
            qkv_bias (bool): If True, add a learnable bias to query, key, value.
            norm_layer (nn.Module): Normalization layer.
            act_layer (nn.Module): Activation layer.
            use_abs_pos (bool): If True, use absolute positional embeddings.
            use_rel_pos (bool): If True, add relative positional embeddings to the attention map.
            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
            window_size (int): Window size for window attention blocks.
            global_attn_indexes (list): Indexes for blocks using global attention.
        """
        super().__init__()
        self.img_size = img_size

        self.patch_embed = PatchEmbed(
            kernel_size=(patch_size, patch_size),
            stride=(patch_size, patch_size),
            in_chans=in_chans,
            embed_dim=embed_dim,
        )

        if use_abs_pos:
            # Initialize absolute positional embedding with pretrain image size.
            self.pos_embed = mx.zeros(
                [1, img_size // patch_size, img_size // patch_size, embed_dim]
            )
        else:
            self.pos_embed = None

        self.layers = []
        for i in range(depth):
            block = Block(
                dim=embed_dim,
                num_heads=num_heads,
                mlp_ratio=mlp_ratio,
                qkv_bias=qkv_bias,
                norm_layer=norm_layer,
                act_layer=act_layer,
                use_rel_pos=use_rel_pos,
                rel_pos_zero_init=rel_pos_zero_init,
                window_size=window_size if i not in global_attn_indexes else 0,
                input_size=(img_size // patch_size, img_size // patch_size),
            )
            self.layers.append(block)

        self.neck = Neck(embed_dim, out_chans)

    def __call__(self, x: mx.array) -> mx.array:
        x = self.patch_embed(x)
        if self.pos_embed is not None:
            x = x + self.pos_embed

        for blk in self.layers:
            x = blk(x)

        x = self.neck(x)
        return x


class Neck(nn.Module):
    def __init__(self, embed_dim, out_chans):
        super().__init__()
        self.conv1 = nn.Conv2d(
            embed_dim,
            out_chans,
            kernel_size=1,
            bias=False,
        )
        self.layer_norm1 = LayerNorm2d(out_chans)
        self.conv2 = nn.Conv2d(
            out_chans,
            out_chans,
            kernel_size=3,
            padding=1,
            bias=False,
        )
        self.layer_norm2 = LayerNorm2d(out_chans)

    def __call__(self, x):
        return self.layer_norm2(self.conv2(self.layer_norm1(self.conv1(x))))


class Block(nn.Module):
    """Transformer blocks with support of window attention and residual propagation blocks"""

    def __init__(
        self,
        dim: int,
        num_heads: int,
        mlp_ratio: float = 4.0,
        qkv_bias: bool = True,
        norm_layer: Type[nn.Module] = nn.LayerNorm,
        act_layer: Type[nn.Module] = nn.GELU,
        use_rel_pos: bool = False,
        rel_pos_zero_init: bool = True,
        window_size: int = 0,
        input_size: Optional[Tuple[int, int]] = None,
    ) -> None:
        """
        Args:
            dim (int): Number of input channels.
            num_heads (int): Number of attention heads in each ViT block.
            mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
            qkv_bias (bool): If True, add a learnable bias to query, key, value.
            norm_layer (nn.Module): Normalization layer.
            act_layer (nn.Module): Activation layer.
            use_rel_pos (bool): If True, add relative positional embeddings to the attention map.
            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
            window_size (int): Window size for window attention blocks. If it equals 0, then
                use global attention.
            input_size (tuple(int, int) or None): Input resolution for calculating the relative
                positional parameter size.
        """
        super().__init__()
        self.layer_norm1 = norm_layer(dim)
        self.attn = Attention(
            dim,
            num_heads=num_heads,
            qkv_bias=qkv_bias,
            use_rel_pos=use_rel_pos,
            rel_pos_zero_init=rel_pos_zero_init,
            input_size=input_size if window_size == 0 else (window_size, window_size),
        )

        self.layer_norm2 = norm_layer(dim)
        self.mlp = MLPBlock(
            embedding_dim=dim, mlp_dim=int(dim * mlp_ratio), act=act_layer
        )

        self.window_size = window_size

    def __call__(self, x: mx.array) -> mx.array:
        shortcut = x
        x = self.layer_norm1(x)
        # Window partition
        if self.window_size > 0:
            H, W = x.shape[1], x.shape[2]
            x, pad_hw = window_partition(x, self.window_size)

        x = self.attn(x)
        # Reverse window partition
        if self.window_size > 0:
            x = window_unpartition(x, self.window_size, pad_hw, (H, W))

        x = shortcut + x
        x = x + self.mlp(self.layer_norm2(x))

        return x


class Attention(nn.Module):
    """Multi-head Attention block with relative position embeddings."""

    def __init__(
        self,
        dim: int,
        num_heads: int = 8,
        qkv_bias: bool = True,
        use_rel_pos: bool = False,
        rel_pos_zero_init: bool = True,
        input_size: Optional[Tuple[int, int]] = None,
    ) -> None:
        """
        Args:
            dim (int): Number of input channels.
            num_heads (int): Number of attention heads.
            qkv_bias (bool):  If True, add a learnable bias to query, key, value.
            rel_pos (bool): If True, add relative positional embeddings to the attention map.
            rel_pos_zero_init (bool): If True, zero initialize relative positional parameters.
            input_size (tuple(int, int) or None): Input resolution for calculating the relative
                positional parameter size.
        """
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // num_heads
        self.scale = head_dim**-0.5

        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
        self.proj = nn.Linear(dim, dim)

        self.use_rel_pos = use_rel_pos
        if self.use_rel_pos:
            assert (
                input_size is not None
            ), "Input size must be provided if using relative positional encoding."
            # initialize relative positional embeddings
            self.rel_pos_h = mx.zeros(shape=(2 * input_size[0] - 1, head_dim))
            self.rel_pos_w = mx.zeros(shape=(2 * input_size[1] - 1, head_dim))

    def __call__(self, x: mx.array) -> mx.array:
        B, H, W, _ = x.shape
        # qkv with shape (3, B, nHead, H * W, C)
        qkv = (
            self.qkv(x)
            .reshape(B, H * W, 3, self.num_heads, -1)
            .transpose(2, 0, 3, 1, 4)
        )

        # q, k, v with shape (B * nHead, H * W, C)
        qkv = qkv.reshape(3, B * self.num_heads, H * W, -1)
        q, k, v = qkv[0], qkv[1], qkv[2]

        attn = (q * self.scale) @ k.transpose(0, 2, 1)

        if self.use_rel_pos:
            attn = add_decomposed_rel_pos(
                attn, q, self.rel_pos_h, self.rel_pos_w, (H, W), (H, W)
            )

        attn = mx.softmax(attn, axis=-1)
        x = (
            (attn @ v)
            .reshape(B, self.num_heads, H, W, -1)
            .transpose(0, 2, 3, 1, 4)
            .reshape(B, H, W, -1)
        )
        x = self.proj(x)

        return x


def window_partition(x: mx.array, window_size: int) -> Tuple[mx.array, Tuple[int, int]]:
    """
    Partition into non-overlapping windows with padding if needed.
    Args:
        x (mx.array): input tokens with [B, H, W, C].
        window_size (int): window size.

    Returns:
        windows: windows after partition with [B * num_windows, window_size, window_size, C].
        (Hp, Wp): padded height and width before partition
    """
    B, H, W, C = x.shape

    pad_h = (window_size - H % window_size) % window_size
    pad_w = (window_size - W % window_size) % window_size
    if pad_h > 0 or pad_w > 0:
        x = mx.pad(x, ((0, 0), (0, pad_w), (0, pad_h), (0, 0)))
    Hp, Wp = H + pad_h, W + pad_w

    x = x.reshape(B, Hp // window_size, window_size, Wp // window_size, window_size, C)
    windows = x.transpose(0, 1, 3, 2, 4, 5).reshape(-1, window_size, window_size, C)
    return windows, (Hp, Wp)


def window_unpartition(
    windows: mx.array,
    window_size: int,
    pad_hw: Tuple[int, int],
    hw: Tuple[int, int],
) -> mx.array:
    """
    Window unpartition into original sequences and removing padding.
    Args:
        windows (mx.array): input tokens with [B * num_windows, window_size, window_size, C].
        window_size (int): window size.
        pad_hw (Tuple): padded height and width (Hp, Wp).
        hw (Tuple): original height and width (H, W) before padding.

    Returns:
        x: unpartitioned sequences with [B, H, W, C].
    """
    Hp, Wp = pad_hw
    H, W = hw
    B = windows.shape[0] // (Hp * Wp // window_size // window_size)
    x = windows.reshape(
        B, Hp // window_size, Wp // window_size, window_size, window_size, -1
    )
    x = x.transpose(0, 1, 3, 2, 4, 5).reshape(B, Hp, Wp, -1)

    if Hp > H or Wp > W:
        x = x[:, :H, :W, :]
    return x


def get_rel_pos(q_size: int, k_size: int, rel_pos: mx.array) -> mx.array:
    """
    Get relative positional embeddings according to the relative positions of
        query and key sizes.
    Args:
        q_size (int): size of query q.
        k_size (int): size of key k.
        rel_pos (mx.array): relative position embeddings (L, C).

    Returns:
        Extracted positional embeddings according to relative positions.
    """
    max_rel_dist = int(2 * max(q_size, k_size) - 1)
    # Interpolate rel pos if needed.
    if rel_pos.shape[0] != max_rel_dist:
        # Interpolate rel pos.
        rel_pos_resized = rel_pos.reshape(1, rel_pos.shape[0], -1).transpose(0, 2, 1)
        scale_factor = (
            max_rel_dist[0] / rel_pos_resized.shape[1],
            max_rel_dist[1] / rel_pos_resized.shape[2],
        )
        rel_pos_resized = nn.Upsample(scale_factor=scale_factor, mode="linear")(
            rel_pos_resized
        )
        rel_pos_resized = rel_pos_resized.reshape(-1, max_rel_dist).transpose(1, 0)
    else:
        rel_pos_resized = rel_pos

    # Scale the coords with short length if shapes for q and k are different.
    q_coords = mx.arange(q_size)[:, None] * max(k_size / q_size, 1.0)
    k_coords = mx.arange(k_size)[None, :] * max(q_size / k_size, 1.0)
    relative_coords = (q_coords - k_coords) + (k_size - 1) * max(q_size / k_size, 1.0)

    return rel_pos_resized[relative_coords.astype(mx.int64)]


def add_decomposed_rel_pos(
    attn,
    q,
    rel_pos_h: mx.array,
    rel_pos_w: mx.array,
    q_size: Tuple[int, int],
    k_size: Tuple[int, int],
) -> mx.array:
    """
    Calculate decomposed Relative Positional Embeddings from :paper:`mvitv2`.
    https://github.com/facebookresearch/mvit/blob/19786631e330df9f3622e5402b4a419a263a2c80/mvit/models/attention.py   # noqa B950
    Args:
        attn (mx.array): attention map.
        q (mx.array): query q in the attention layer with shape (B, q_h * q_w, C).
        rel_pos_h (mx.array): relative position embeddings (Lh, C) for height axis.
        rel_pos_w (mx.array): relative position embeddings (Lw, C) for width axis.
        q_size (Tuple): spatial sequence size of query q with (q_h, q_w).
        k_size (Tuple): spatial sequence size of key k with (k_h, k_w).

    Returns:
        attn (mx.array): attention map with added relative positional embeddings.
    """
    q_h, q_w = q_size
    k_h, k_w = k_size
    Rh = get_rel_pos(q_h, k_h, rel_pos_h)
    Rw = get_rel_pos(q_w, k_w, rel_pos_w)

    B, _, dim = q.shape
    r_q = q.reshape(B, q_h, q_w, dim)

    # TODO: replace mx.einsum when its ready
    # workaround for these einsum computations
    # rel_h = torch.einsum("bhwc,hkc->bhwk", r_q, Rh)
    # rel_w = torch.einsum("bhwc,wkc->bhwk", r_q, Rw)
    rel_h = r_q @ Rh.transpose(0, 2, 1)
    rel_w = (r_q.transpose(0, 2, 1, 3) @ Rw.transpose(0, 2, 1)).transpose(0, 2, 1, 3)

    attn = (
        attn.reshape(B, q_h, q_w, k_h, k_w)
        + rel_h[:, :, :, :, None]
        + rel_w[:, :, :, None, :]
    ).reshape(B, q_h * q_w, k_h * k_w)

    return attn


class PatchEmbed(nn.Module):
    """
    Image to Patch Embedding.
    """

    def __init__(
        self,
        kernel_size: Tuple[int, int] = (16, 16),
        stride: Tuple[int, int] = (16, 16),
        padding: Tuple[int, int] = (0, 0),
        in_chans: int = 3,
        embed_dim: int = 768,
    ) -> None:
        """
        Args:
            kernel_size (Tuple): kernel size of the projection layer.
            stride (Tuple): stride of the projection layer.
            padding (Tuple): padding size of the projection layer.
            in_chans (int): Number of input image channels.
            embed_dim (int): Patch embedding dimension.
        """
        super().__init__()

        self.projection = nn.Conv2d(
            in_chans, embed_dim, kernel_size=kernel_size, stride=stride, padding=padding
        )

    def __call__(self, x: mx.array) -> mx.array:
        x = self.projection(x)
        return x