─░├žeri─če ge├ž

Referans i├žin ultralytics/models/sam/modules/transformer.py

Not

Bu dosya https://github.com/ultralytics/ultralytics/blob/main/ ultralytics/models/ sam/modules/transformer .py adresinde mevcuttur. Bir sorun tespit ederseniz l├╝tfen bir ├çekme ─░ste─či ­čŤá´ŞĆ ile katk─▒da bulunarak d├╝zeltilmesine yard─▒mc─▒ olun. Te┼čekk├╝rler ­čÖĆ!



ultralytics.models.sam.modules.transformer.TwoWayTransformer

├ťsler: Module

Hem g├Âr├╝nt├╝ hem de sorgu noktalar─▒na ayn─▒ anda dikkat edilmesini sa─člayan ─░ki Y├Ânl├╝ D├Ân├╝┼čt├╝r├╝c├╝ mod├╝l├╝. Bu s─▒n─▒f konumsal g├Âm├╝l├╝ sorgular─▒ kullanarak bir giri┼č g├Âr├╝nt├╝s├╝ne kat─▒lan ├Âzel bir d├Ân├╝┼čt├╝r├╝c├╝ kod ├ž├Âz├╝c├╝ olarak hizmet eder sa─član─▒r. Bu ├Âzellikle nesne alg─▒lama, g├Âr├╝nt├╝ segmentasyonu ve nokta bulutu gibi g├Ârevler i├žin kullan─▒┼čl─▒d─▒r i┼členiyor.

Nitelikler:

─░sim Tip A├ž─▒klama
depth int

Transformat├Ârdeki katman say─▒s─▒.

embedding_dim int

Giri┼č kat─▒┼čt─▒rmalar─▒ i├žin kanal boyutu.

num_heads int

├çok kafal─▒ dikkat i├žin kafa say─▒s─▒.

mlp_dim int

MLP blo─ču i├žin dahili kanal boyutu.

layers ModuleList

D├Ân├╝┼čt├╝r├╝c├╝y├╝ olu┼čturan TwoWayAttentionBlock katmanlar─▒n─▒n listesi.

final_attn_token_to_image Attention

Sorgulardan g├Âr├╝nt├╝ye uygulanan son dikkat katman─▒.

norm_final_attn LayerNorm

Nihai sorgulara uygulanan katman normalizasyonu.

Kaynak kodu ultralytics/models/sam/modules/transformer.py
class TwoWayTransformer(nn.Module):
    """
    A Two-Way Transformer module that enables the simultaneous attention to both image and query points. This class
    serves as a specialized transformer decoder that attends to an input image using queries whose positional embedding
    is supplied. This is particularly useful for tasks like object detection, image segmentation, and point cloud
    processing.

    Attributes:
        depth (int): The number of layers in the transformer.
        embedding_dim (int): The channel dimension for the input embeddings.
        num_heads (int): The number of heads for multihead attention.
        mlp_dim (int): The internal channel dimension for the MLP block.
        layers (nn.ModuleList): The list of TwoWayAttentionBlock layers that make up the transformer.
        final_attn_token_to_image (Attention): The final attention layer applied from the queries to the image.
        norm_final_attn (nn.LayerNorm): The layer normalization applied to the final queries.
    """

    def __init__(
        self,
        depth: int,
        embedding_dim: int,
        num_heads: int,
        mlp_dim: int,
        activation: Type[nn.Module] = nn.ReLU,
        attention_downsample_rate: int = 2,
    ) -> None:
        """
        A transformer decoder that attends to an input image using queries whose positional embedding is supplied.

        Args:
          depth (int): number of layers in the transformer
          embedding_dim (int): the channel dimension for the input embeddings
          num_heads (int): the number of heads for multihead attention. Must
            divide embedding_dim
          mlp_dim (int): the channel dimension internal to the MLP block
          activation (nn.Module): the activation to use in the MLP block
        """
        super().__init__()
        self.depth = depth
        self.embedding_dim = embedding_dim
        self.num_heads = num_heads
        self.mlp_dim = mlp_dim
        self.layers = nn.ModuleList()

        for i in range(depth):
            self.layers.append(
                TwoWayAttentionBlock(
                    embedding_dim=embedding_dim,
                    num_heads=num_heads,
                    mlp_dim=mlp_dim,
                    activation=activation,
                    attention_downsample_rate=attention_downsample_rate,
                    skip_first_layer_pe=(i == 0),
                )
            )

        self.final_attn_token_to_image = Attention(embedding_dim, num_heads, downsample_rate=attention_downsample_rate)
        self.norm_final_attn = nn.LayerNorm(embedding_dim)

    def forward(
        self,
        image_embedding: Tensor,
        image_pe: Tensor,
        point_embedding: Tensor,
    ) -> Tuple[Tensor, Tensor]:
        """
        Args:
          image_embedding (torch.Tensor): image to attend to. Should be shape B x embedding_dim x h x w for any h and w.
          image_pe (torch.Tensor): the positional encoding to add to the image. Must have same shape as image_embedding.
          point_embedding (torch.Tensor): the embedding to add to the query points.
            Must have shape B x N_points x embedding_dim for any N_points.

        Returns:
          (torch.Tensor): the processed point_embedding
          (torch.Tensor): the processed image_embedding
        """
        # BxCxHxW -> BxHWxC == B x N_image_tokens x C
        bs, c, h, w = image_embedding.shape
        image_embedding = image_embedding.flatten(2).permute(0, 2, 1)
        image_pe = image_pe.flatten(2).permute(0, 2, 1)

        # Prepare queries
        queries = point_embedding
        keys = image_embedding

        # Apply transformer blocks and final layernorm
        for layer in self.layers:
            queries, keys = layer(
                queries=queries,
                keys=keys,
                query_pe=point_embedding,
                key_pe=image_pe,
            )

        # Apply the final attention layer from the points to the image
        q = queries + point_embedding
        k = keys + image_pe
        attn_out = self.final_attn_token_to_image(q=q, k=k, v=keys)
        queries = queries + attn_out
        queries = self.norm_final_attn(queries)

        return queries, keys

__init__(depth, embedding_dim, num_heads, mlp_dim, activation=nn.ReLU, attention_downsample_rate=2)

Konumsal g├Âmme sa─članan sorgular─▒ kullanarak bir giri┼č g├Âr├╝nt├╝s├╝ne kat─▒lan bir d├Ân├╝┼čt├╝r├╝c├╝ kod ├ž├Âz├╝c├╝.

Parametreler:

─░sim Tip A├ž─▒klama Varsay─▒lan
depth int

transformat├Ârdeki katman say─▒s─▒

gerekli
embedding_dim int

giri┼č kat─▒┼čt─▒rmalar─▒ i├žin kanal boyutu

gerekli
num_heads int

├žok kafal─▒ dikkat i├žin kafa say─▒s─▒. Zorunlu embedding_dim'i b├Âl

gerekli
mlp_dim int

MLP blo─čuna dahili kanal boyutu

gerekli
activation Module

MLP blo─čunda kullan─▒lacak aktivasyon

ReLU
Kaynak kodu ultralytics/models/sam/modules/transformer.py
def __init__(
    self,
    depth: int,
    embedding_dim: int,
    num_heads: int,
    mlp_dim: int,
    activation: Type[nn.Module] = nn.ReLU,
    attention_downsample_rate: int = 2,
) -> None:
    """
    A transformer decoder that attends to an input image using queries whose positional embedding is supplied.

    Args:
      depth (int): number of layers in the transformer
      embedding_dim (int): the channel dimension for the input embeddings
      num_heads (int): the number of heads for multihead attention. Must
        divide embedding_dim
      mlp_dim (int): the channel dimension internal to the MLP block
      activation (nn.Module): the activation to use in the MLP block
    """
    super().__init__()
    self.depth = depth
    self.embedding_dim = embedding_dim
    self.num_heads = num_heads
    self.mlp_dim = mlp_dim
    self.layers = nn.ModuleList()

    for i in range(depth):
        self.layers.append(
            TwoWayAttentionBlock(
                embedding_dim=embedding_dim,
                num_heads=num_heads,
                mlp_dim=mlp_dim,
                activation=activation,
                attention_downsample_rate=attention_downsample_rate,
                skip_first_layer_pe=(i == 0),
            )
        )

    self.final_attn_token_to_image = Attention(embedding_dim, num_heads, downsample_rate=attention_downsample_rate)
    self.norm_final_attn = nn.LayerNorm(embedding_dim)

forward(image_embedding, image_pe, point_embedding)

Parametreler:

─░sim Tip A├ž─▒klama Varsay─▒lan
image_embedding Tensor

ilgilenilecek g├Âr├╝nt├╝. Herhangi bir h ve w i├žin B x embedding_dim x h x w ┼čeklinde olmal─▒d─▒r.

gerekli
image_pe Tensor

g├Âr├╝nt├╝ye eklenecek konumsal kodlama. image_embedding ile ayn─▒ ┼čekle sahip olmal─▒d─▒r.

gerekli
point_embedding Tensor

sorgu noktalar─▒na eklenecek g├Âmme. Herhangi bir N_points i├žin B x N_points x embedding_dim ┼čekline sahip olmal─▒d─▒r.

gerekli

─░ade:

Tip A├ž─▒klama
Tensor

i┼členmi┼č point_embedding

Tensor

i┼členmi┼č g├Âr├╝nt├╝_embedding

Kaynak kodu ultralytics/models/sam/modules/transformer.py
def forward(
    self,
    image_embedding: Tensor,
    image_pe: Tensor,
    point_embedding: Tensor,
) -> Tuple[Tensor, Tensor]:
    """
    Args:
      image_embedding (torch.Tensor): image to attend to. Should be shape B x embedding_dim x h x w for any h and w.
      image_pe (torch.Tensor): the positional encoding to add to the image. Must have same shape as image_embedding.
      point_embedding (torch.Tensor): the embedding to add to the query points.
        Must have shape B x N_points x embedding_dim for any N_points.

    Returns:
      (torch.Tensor): the processed point_embedding
      (torch.Tensor): the processed image_embedding
    """
    # BxCxHxW -> BxHWxC == B x N_image_tokens x C
    bs, c, h, w = image_embedding.shape
    image_embedding = image_embedding.flatten(2).permute(0, 2, 1)
    image_pe = image_pe.flatten(2).permute(0, 2, 1)

    # Prepare queries
    queries = point_embedding
    keys = image_embedding

    # Apply transformer blocks and final layernorm
    for layer in self.layers:
        queries, keys = layer(
            queries=queries,
            keys=keys,
            query_pe=point_embedding,
            key_pe=image_pe,
        )

    # Apply the final attention layer from the points to the image
    q = queries + point_embedding
    k = keys + image_pe
    attn_out = self.final_attn_token_to_image(q=q, k=k, v=keys)
    queries = queries + attn_out
    queries = self.norm_final_attn(queries)

    return queries, keys



ultralytics.models.sam.modules.transformer.TwoWayAttentionBlock

├ťsler: Module

Hem kendi kendine dikkat hem de ├žapraz dikkati iki y├Ânde ger├žekle┼čtiren bir dikkat blo─ču: anahtarlara ve sorgular i├žin anahtarlar. Bu blok d├Ârt ana katmandan olu┼čur: (1) seyrek girdiler ├╝zerinde ├Âz dikkat, (2) ├žapraz dikkat seyrek girdilerin yo─čun girdilere, (3) seyrek girdiler ├╝zerinde bir MLP blo─ču ve (4) yo─čun girdilerin seyrek girdiler.

Nitelikler:

─░sim Tip A├ž─▒klama
self_attn Attention

Sorgular i├žin ├Âz dikkat katman─▒.

norm1 LayerNorm

─░lk dikkat blo─čunun ard─▒ndan katman normalizasyonu.

cross_attn_token_to_image Attention

Sorgulardan anahtarlara ├žapraz dikkat katman─▒.

norm2 LayerNorm

─░kinci dikkat blo─čunun ard─▒ndan katman normalizasyonu.

mlp MLPBlock

Sorgu g├Âm├╝lerini d├Ân├╝┼čt├╝ren MLP blo─ču.

norm3 LayerNorm

MLP blo─čunu takip eden katman normalizasyonu.

norm4 LayerNorm

├ť├ž├╝nc├╝ dikkat blo─čunun ard─▒ndan katman normalizasyonu.

cross_attn_image_to_token Attention

Anahtarlardan sorgulara ├žapraz dikkat katman─▒.

skip_first_layer_pe bool

─░lk katmandaki konumsal kodlaman─▒n atlan─▒p atlanmayaca─č─▒.

Kaynak kodu ultralytics/models/sam/modules/transformer.py
class TwoWayAttentionBlock(nn.Module):
    """
    An attention block that performs both self-attention and cross-attention in two directions: queries to keys and
    keys to queries. This block consists of four main layers: (1) self-attention on sparse inputs, (2) cross-attention
    of sparse inputs to dense inputs, (3) an MLP block on sparse inputs, and (4) cross-attention of dense inputs to
    sparse inputs.

    Attributes:
        self_attn (Attention): The self-attention layer for the queries.
        norm1 (nn.LayerNorm): Layer normalization following the first attention block.
        cross_attn_token_to_image (Attention): Cross-attention layer from queries to keys.
        norm2 (nn.LayerNorm): Layer normalization following the second attention block.
        mlp (MLPBlock): MLP block that transforms the query embeddings.
        norm3 (nn.LayerNorm): Layer normalization following the MLP block.
        norm4 (nn.LayerNorm): Layer normalization following the third attention block.
        cross_attn_image_to_token (Attention): Cross-attention layer from keys to queries.
        skip_first_layer_pe (bool): Whether to skip the positional encoding in the first layer.
    """

    def __init__(
        self,
        embedding_dim: int,
        num_heads: int,
        mlp_dim: int = 2048,
        activation: Type[nn.Module] = nn.ReLU,
        attention_downsample_rate: int = 2,
        skip_first_layer_pe: bool = False,
    ) -> None:
        """
        A transformer block with four layers: (1) self-attention of sparse inputs, (2) cross attention of sparse
        inputs to dense inputs, (3) mlp block on sparse inputs, and (4) cross attention of dense inputs to sparse
        inputs.

        Args:
          embedding_dim (int): the channel dimension of the embeddings
          num_heads (int): the number of heads in the attention layers
          mlp_dim (int): the hidden dimension of the mlp block
          activation (nn.Module): the activation of the mlp block
          skip_first_layer_pe (bool): skip the PE on the first layer
        """
        super().__init__()
        self.self_attn = Attention(embedding_dim, num_heads)
        self.norm1 = nn.LayerNorm(embedding_dim)

        self.cross_attn_token_to_image = Attention(embedding_dim, num_heads, downsample_rate=attention_downsample_rate)
        self.norm2 = nn.LayerNorm(embedding_dim)

        self.mlp = MLPBlock(embedding_dim, mlp_dim, activation)
        self.norm3 = nn.LayerNorm(embedding_dim)

        self.norm4 = nn.LayerNorm(embedding_dim)
        self.cross_attn_image_to_token = Attention(embedding_dim, num_heads, downsample_rate=attention_downsample_rate)

        self.skip_first_layer_pe = skip_first_layer_pe

    def forward(self, queries: Tensor, keys: Tensor, query_pe: Tensor, key_pe: Tensor) -> Tuple[Tensor, Tensor]:
        """Apply self-attention and cross-attention to queries and keys and return the processed embeddings."""

        # Self attention block
        if self.skip_first_layer_pe:
            queries = self.self_attn(q=queries, k=queries, v=queries)
        else:
            q = queries + query_pe
            attn_out = self.self_attn(q=q, k=q, v=queries)
            queries = queries + attn_out
        queries = self.norm1(queries)

        # Cross attention block, tokens attending to image embedding
        q = queries + query_pe
        k = keys + key_pe
        attn_out = self.cross_attn_token_to_image(q=q, k=k, v=keys)
        queries = queries + attn_out
        queries = self.norm2(queries)

        # MLP block
        mlp_out = self.mlp(queries)
        queries = queries + mlp_out
        queries = self.norm3(queries)

        # Cross attention block, image embedding attending to tokens
        q = queries + query_pe
        k = keys + key_pe
        attn_out = self.cross_attn_image_to_token(q=k, k=q, v=queries)
        keys = keys + attn_out
        keys = self.norm4(keys)

        return queries, keys

__init__(embedding_dim, num_heads, mlp_dim=2048, activation=nn.ReLU, attention_downsample_rate=2, skip_first_layer_pe=False)

D├Ârt katmanl─▒ bir d├Ân├╝┼čt├╝r├╝c├╝ blo─ču: (1) seyrek girdilerin kendi kendine dikkati, (2) seyrek girdilerin ├žapraz dikkati yo─čun girdilere, (3) seyrek girdiler ├╝zerinde mlp blo─ču ve (4) yo─čun girdilerin seyrek girdilere ├žapraz dikkati Girdiler.

Parametreler:

─░sim Tip A├ž─▒klama Varsay─▒lan
embedding_dim int

g├Âm├╝lmelerin kanal boyutu

gerekli
num_heads int

dikkat katmanlar─▒ndaki kafa say─▒s─▒

gerekli
mlp_dim int

mlp blo─čunun gi╠çzli╠ç boyutu

2048
activation Module

mlp blo─čunun akti╠çvasyonu

ReLU
skip_first_layer_pe bool

ilk katmandaki PE'yi atlay─▒n

False
Kaynak kodu ultralytics/models/sam/modules/transformer.py
def __init__(
    self,
    embedding_dim: int,
    num_heads: int,
    mlp_dim: int = 2048,
    activation: Type[nn.Module] = nn.ReLU,
    attention_downsample_rate: int = 2,
    skip_first_layer_pe: bool = False,
) -> None:
    """
    A transformer block with four layers: (1) self-attention of sparse inputs, (2) cross attention of sparse
    inputs to dense inputs, (3) mlp block on sparse inputs, and (4) cross attention of dense inputs to sparse
    inputs.

    Args:
      embedding_dim (int): the channel dimension of the embeddings
      num_heads (int): the number of heads in the attention layers
      mlp_dim (int): the hidden dimension of the mlp block
      activation (nn.Module): the activation of the mlp block
      skip_first_layer_pe (bool): skip the PE on the first layer
    """
    super().__init__()
    self.self_attn = Attention(embedding_dim, num_heads)
    self.norm1 = nn.LayerNorm(embedding_dim)

    self.cross_attn_token_to_image = Attention(embedding_dim, num_heads, downsample_rate=attention_downsample_rate)
    self.norm2 = nn.LayerNorm(embedding_dim)

    self.mlp = MLPBlock(embedding_dim, mlp_dim, activation)
    self.norm3 = nn.LayerNorm(embedding_dim)

    self.norm4 = nn.LayerNorm(embedding_dim)
    self.cross_attn_image_to_token = Attention(embedding_dim, num_heads, downsample_rate=attention_downsample_rate)

    self.skip_first_layer_pe = skip_first_layer_pe

forward(queries, keys, query_pe, key_pe)

Sorgulara ve anahtarlara ├Âz dikkat ve ├žapraz dikkat uygulay─▒n ve i┼členmi┼č kat─▒┼čt─▒rmalar─▒ d├Ând├╝r├╝n.

Kaynak kodu ultralytics/models/sam/modules/transformer.py
def forward(self, queries: Tensor, keys: Tensor, query_pe: Tensor, key_pe: Tensor) -> Tuple[Tensor, Tensor]:
    """Apply self-attention and cross-attention to queries and keys and return the processed embeddings."""

    # Self attention block
    if self.skip_first_layer_pe:
        queries = self.self_attn(q=queries, k=queries, v=queries)
    else:
        q = queries + query_pe
        attn_out = self.self_attn(q=q, k=q, v=queries)
        queries = queries + attn_out
    queries = self.norm1(queries)

    # Cross attention block, tokens attending to image embedding
    q = queries + query_pe
    k = keys + key_pe
    attn_out = self.cross_attn_token_to_image(q=q, k=k, v=keys)
    queries = queries + attn_out
    queries = self.norm2(queries)

    # MLP block
    mlp_out = self.mlp(queries)
    queries = queries + mlp_out
    queries = self.norm3(queries)

    # Cross attention block, image embedding attending to tokens
    q = queries + query_pe
    k = keys + key_pe
    attn_out = self.cross_attn_image_to_token(q=k, k=q, v=queries)
    keys = keys + attn_out
    keys = self.norm4(keys)

    return queries, keys



ultralytics.models.sam.modules.transformer.Attention

├ťsler: Module

Sorgulara, anahtarlara ve projeksiyondan sonra g├Âmme boyutunun k├╝├ž├╝lt├╝lmesine izin veren bir dikkat katman─▒ de─čerler.

Kaynak kodu ultralytics/models/sam/modules/transformer.py
class Attention(nn.Module):
    """An attention layer that allows for downscaling the size of the embedding after projection to queries, keys, and
    values.
    """

    def __init__(
        self,
        embedding_dim: int,
        num_heads: int,
        downsample_rate: int = 1,
    ) -> None:
        """
        Initializes the Attention model with the given dimensions and settings.

        Args:
            embedding_dim (int): The dimensionality of the input embeddings.
            num_heads (int): The number of attention heads.
            downsample_rate (int, optional): The factor by which the internal dimensions are downsampled. Defaults to 1.

        Raises:
            AssertionError: If 'num_heads' does not evenly divide the internal dim (embedding_dim / downsample_rate).
        """
        super().__init__()
        self.embedding_dim = embedding_dim
        self.internal_dim = embedding_dim // downsample_rate
        self.num_heads = num_heads
        assert self.internal_dim % num_heads == 0, "num_heads must divide embedding_dim."

        self.q_proj = nn.Linear(embedding_dim, self.internal_dim)
        self.k_proj = nn.Linear(embedding_dim, self.internal_dim)
        self.v_proj = nn.Linear(embedding_dim, self.internal_dim)
        self.out_proj = nn.Linear(self.internal_dim, embedding_dim)

    @staticmethod
    def _separate_heads(x: Tensor, num_heads: int) -> Tensor:
        """Separate the input tensor into the specified number of attention heads."""
        b, n, c = x.shape
        x = x.reshape(b, n, num_heads, c // num_heads)
        return x.transpose(1, 2)  # B x N_heads x N_tokens x C_per_head

    @staticmethod
    def _recombine_heads(x: Tensor) -> Tensor:
        """Recombine the separated attention heads into a single tensor."""
        b, n_heads, n_tokens, c_per_head = x.shape
        x = x.transpose(1, 2)
        return x.reshape(b, n_tokens, n_heads * c_per_head)  # B x N_tokens x C

    def forward(self, q: Tensor, k: Tensor, v: Tensor) -> Tensor:
        """Compute the attention output given the input query, key, and value tensors."""

        # Input projections
        q = self.q_proj(q)
        k = self.k_proj(k)
        v = self.v_proj(v)

        # Separate into heads
        q = self._separate_heads(q, self.num_heads)
        k = self._separate_heads(k, self.num_heads)
        v = self._separate_heads(v, self.num_heads)

        # Attention
        _, _, _, c_per_head = q.shape
        attn = q @ k.permute(0, 1, 3, 2)  # B x N_heads x N_tokens x N_tokens
        attn = attn / math.sqrt(c_per_head)
        attn = torch.softmax(attn, dim=-1)

        # Get output
        out = attn @ v
        out = self._recombine_heads(out)
        return self.out_proj(out)

__init__(embedding_dim, num_heads, downsample_rate=1)

Attention modelini verilen boyutlar ve ayarlarla ba┼člat─▒r.

Parametreler:

─░sim Tip A├ž─▒klama Varsay─▒lan
embedding_dim int

Giri┼č kat─▒┼čt─▒rmalar─▒n─▒n boyutlulu─ču.

gerekli
num_heads int

Dikkat kafalar─▒n─▒n say─▒s─▒.

gerekli
downsample_rate int

─░├ž boyutlar─▒n k├╝├ž├╝lt├╝ld├╝─č├╝ fakt├Âr. Varsay─▒lan de─čer 1'dir.

1

Zamlar:

Tip A├ž─▒klama
AssertionError

'num_heads' dahili boyutu (embedding_dim / downsample_rate) e┼čit olarak b├Âlmezse.

Kaynak kodu ultralytics/models/sam/modules/transformer.py
def __init__(
    self,
    embedding_dim: int,
    num_heads: int,
    downsample_rate: int = 1,
) -> None:
    """
    Initializes the Attention model with the given dimensions and settings.

    Args:
        embedding_dim (int): The dimensionality of the input embeddings.
        num_heads (int): The number of attention heads.
        downsample_rate (int, optional): The factor by which the internal dimensions are downsampled. Defaults to 1.

    Raises:
        AssertionError: If 'num_heads' does not evenly divide the internal dim (embedding_dim / downsample_rate).
    """
    super().__init__()
    self.embedding_dim = embedding_dim
    self.internal_dim = embedding_dim // downsample_rate
    self.num_heads = num_heads
    assert self.internal_dim % num_heads == 0, "num_heads must divide embedding_dim."

    self.q_proj = nn.Linear(embedding_dim, self.internal_dim)
    self.k_proj = nn.Linear(embedding_dim, self.internal_dim)
    self.v_proj = nn.Linear(embedding_dim, self.internal_dim)
    self.out_proj = nn.Linear(self.internal_dim, embedding_dim)

forward(q, k, v)

Giri┼č sorgusu, anahtar ve de─čer tens├Ârleri verildi─činde dikkat ├ž─▒kt─▒s─▒n─▒ hesaplay─▒n.

Kaynak kodu ultralytics/models/sam/modules/transformer.py
def forward(self, q: Tensor, k: Tensor, v: Tensor) -> Tensor:
    """Compute the attention output given the input query, key, and value tensors."""

    # Input projections
    q = self.q_proj(q)
    k = self.k_proj(k)
    v = self.v_proj(v)

    # Separate into heads
    q = self._separate_heads(q, self.num_heads)
    k = self._separate_heads(k, self.num_heads)
    v = self._separate_heads(v, self.num_heads)

    # Attention
    _, _, _, c_per_head = q.shape
    attn = q @ k.permute(0, 1, 3, 2)  # B x N_heads x N_tokens x N_tokens
    attn = attn / math.sqrt(c_per_head)
    attn = torch.softmax(attn, dim=-1)

    # Get output
    out = attn @ v
    out = self._recombine_heads(out)
    return self.out_proj(out)





Created 2023-11-12, Updated 2024-06-02
Authors: glenn-jocher (5), Burhan-Q (1), Laughing-q (1)