model/unet_vit.py

# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.

# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.


import torch
import torch.nn as nn
from einops import rearrange
from functools import partial
import torch.nn.functional as F
from typing import Optional, Tuple, Type
import timm


class vit_timm(nn.Module):
    def __init__(
        self,
        patch_dim,
        pretrained = False,
    ):
        super(vit_timm, self).__init__()
        self.patch_dim = patch_dim
        self.model = timm.create_model(f'vit_base_patch16_224', num_classes=0)
        if pretrained : 
            # load pretrained imagenet weight
            weight = torch.load('/home/caduser/KOTORI/hoang_graph_matching_v1/vit_base_imagenet.pth', map_location = 'cpu')
            self.model.load_state_dict(weight, strict = False)
        
    def forward(self,x):
#         x = self.model.patch_embed(x)   # remove some layer at this step 
#         x = self.model.pos_drop(x)
        if len(x.shape) == 4 :  # 4, 14, 14, 768
            # orignal shape : bs, patch_dim , patch_dim , embedding_dim
            x = rearrange(x, 'b p1 p2 d -> b (p1 p2) d') 
        x = self.model.blocks(x)
        x = self.model.norm(x)
        x = rearrange(x, 'b (p1 p2) c -> b c p1 p2', p1 = self.patch_dim, p2 = self.patch_dim)
        return x  


class vit_encoder_b(nn.Module):
    def __init__(self):
        super().__init__()
        
        prompt_embed_dim = 256
        image_size = 1024
        vit_patch_size = 16
        image_embedding_size = image_size // vit_patch_size
        encoder_embed_dim=768
        encoder_depth=12
        encoder_num_heads=12
        encoder_global_attn_indexes=[2, 5, 8, 11]
        
        self.model =ImageEncoderViT(
                depth=encoder_depth,
                embed_dim=encoder_embed_dim,
                img_size=image_size,
                mlp_ratio=4,
                norm_layer=partial(torch.nn.LayerNorm, eps=1e-6),
                num_heads=encoder_num_heads,
                patch_size=vit_patch_size,
                qkv_bias=True,
                use_rel_pos=True,
                use_abs_pos = False,
                global_attn_indexes=encoder_global_attn_indexes,
                window_size=14,
                out_chans=prompt_embed_dim,
            )
    def forward(self, x):
        return self.model(x)
    
    
class EncoderBottleneck(nn.Module):
    def __init__(self, in_channels, out_channels, stride=1, base_width=64):
        super().__init__()

        self.downsample = nn.Sequential(
            nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=stride, bias=False),
            nn.BatchNorm2d(out_channels)
        )

        width = int(out_channels * (base_width / 64))

        self.conv1 = nn.Conv2d(in_channels, width, kernel_size=1, stride=1, bias=False)
        self.norm1 = nn.BatchNorm2d(width)

        self.conv2 = nn.Conv2d(width, width, kernel_size=3, stride=2, groups=1, padding=1, dilation=1, bias=False)
        self.norm2 = nn.BatchNorm2d(width)

        self.conv3 = nn.Conv2d(width, out_channels, kernel_size=1, stride=1, bias=False)
        self.norm3 = nn.BatchNorm2d(out_channels)

        self.relu = nn.ReLU(inplace=True)

    def forward(self, x):
        x_down = self.downsample(x)

        x = self.conv1(x)
        x = self.norm1(x)
        x = self.relu(x)

        x = self.conv2(x)
        x = self.norm2(x)
        x = self.relu(x)

        x = self.conv3(x)
        x = self.norm3(x)
        x = x + x_down
        x = self.relu(x)

        return x


class DecoderBottleneck(nn.Module):
    def __init__(self, in_channels, out_channels, scale_factor=2):
        super().__init__()

        self.upsample = nn.Upsample(scale_factor=scale_factor, mode='bilinear', align_corners=True)
        self.layer = nn.Sequential(
            nn.Conv2d(in_channels, out_channels, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_channels),
            nn.ReLU(inplace=True),
            nn.Conv2d(out_channels, out_channels, kernel_size=3, stride=1, padding=1),
            nn.BatchNorm2d(out_channels),
            nn.ReLU(inplace=True)
        )

    def forward(self, x, x_concat=None):
        x = self.upsample(x)

        if x_concat is not None:
#             print(x.shape, x_concat.shape)
            x = torch.cat([x_concat, x], dim=1)

        x = self.layer(x)
        return x

    
def load_weight_for_vit_encoder(pretrained):
    
    weight = None     
    elif pretrained == 'lvm-med-vit':
        path = './checkpoints/lvmmed_vit.torch'
        print(f'Pretrained path of LVM-MED-VIT :  {path}')
        weight = torch.load(path, map_location ='cpu')
        print(f'Number of params in original checkpoint : {len(weight)}')
        for key in list(weight.keys()):
            weight['model.' + key] = weight[key]
            del weight[key]
            
    #print(f'Number of params in final checkpoint : {len(weight)}')
    return weight
        
    

class Encoder(nn.Module):
    def __init__(self, in_channels, out_channels, pretrained, patch_dim):
        super().__init__()

        self.conv1 = nn.Conv2d(in_channels, out_channels, kernel_size=7, stride=2, padding=3, bias=False)
        self.norm1 = nn.BatchNorm2d(out_channels)
        self.relu = nn.ReLU(inplace=True)

        self.encoder1 = EncoderBottleneck(out_channels, out_channels * 2, stride=2)
        self.encoder2 = EncoderBottleneck(out_channels * 2, out_channels * 4, stride=2)
        self.encoder3 = EncoderBottleneck(out_channels * 4, out_channels * 8, stride=2)
#         self.vit = vit_encoder_b() 
        
        
#         if pretrained ==  'scratch' :    
#             print('LOAD VIT RANDOM SUCCESSFULLY')
#             self.vit = vit_timm(pretrained = False) 
#         else :
        if pretrained == 'flava':
            from flava_model import Flava_encoder
            print('LOAD FLAVA MODEL SUCCESSFULLY')
            self.vit = Flava_encoder(patch_dim)
        elif pretrained == 'default' :
            print('LOAD VIT RANDOM SUCCESSFULLY')
            self.vit = vit_timm(pretrained = False) 
        elif pretrained == 'clip' :
            from clip_model import Clip_encoder
            print('LOAD CLIP MODEL SUCCESSFULLY')
            self.vit = Clip_encoder(patch_dim)
        elif pretrained == 'imagenet':
            print('LOAD VIT IMAGENET SUCCESSFULLY')
            self.vit = vit_timm(pretrained = True, patch_dim = patch_dim) 
        elif pretrained in  ['sam' , 'ssl' , 'ssl_large']:
            self.vit = vit_encoder_b() 
            weight = load_weight_for_vit_encoder(pretrained)
            self.vit.load_state_dict(weight,strict = False)

        
        self.conv2 = nn.Conv2d(out_channels * 8, 384, kernel_size=3, stride=1, padding=1)
        self.norm2 = nn.BatchNorm2d(384)

    def forward(self, x):
        x = self.conv1(x)
        x = self.norm1(x)
        x1 = self.relu(x)  # 4, 96, 112, 112
   
        x2 = self.encoder1(x1)  # 4, 192, 56, 56
        x3 = self.encoder2(x2)  # 4, 384, 28, 28
        x = self.encoder3(x3)  # 4, 768, 14, 14
#         x = self.conv1x1(x)
        
        x = x.permute(0, 2, 3, 1)    # 4, 14, 14, 768
        
        '''
        enter VIT at this step
        VIT in TranUNET need input shape of (bs, patch_1, patch_2, dim) ie,: 4, 14, 14, 768
                             output shape of (bs, dim, patch_1, patch_2): ie 4, 768, 14, 14
        '''
        x = self.vit(x) # 4, 768, 14, 14
        x = self.conv2(x) # 4, 512, 28, 28 
        x = self.norm2(x)
        x = self.relu(x)
        return x, x1, x2, x3


class Decoder(nn.Module):
    def __init__(self, out_channels, class_num):
        super().__init__()

        self.decoder1 = DecoderBottleneck(out_channels * 8, out_channels * 2)
        self.decoder2 = DecoderBottleneck(out_channels * 4, out_channels)
        self.decoder3 = DecoderBottleneck(out_channels * 2, int(out_channels * 1 / 2))
        self.decoder4 = DecoderBottleneck(int(out_channels * 1 / 2), int(out_channels * 1 / 8))
        self.conv1 = nn.Conv2d(int(out_channels * 1 / 8), class_num, kernel_size=1)

    def forward(self, x, x1, x2, x3):
        x = self.decoder1(x, x3)
        x = self.decoder2(x, x2)
        x = self.decoder3(x, x1)
        x = self.decoder4(x)
        x = self.conv1(x)

        return x


class TransUNet(nn.Module):
    def __init__(self, in_channels, out_channels, class_num, pretrained, patch_dim):
        super().__init__()

        self.encoder = Encoder(in_channels, out_channels, pretrained, patch_dim)
        self.decoder = Decoder(out_channels, class_num)

    def forward(self, x):
        x, x1, x2, x3 = self.encoder(x)
        x = self.decoder(x, x1, x2, x3)
        return x
        


class MLPBlock(nn.Module):
    def __init__(
        self,
        embedding_dim: int,
        mlp_dim: int,
        act: Type[nn.Module] = nn.GELU,
    ) -> None:
        super().__init__()
        self.lin1 = nn.Linear(embedding_dim, mlp_dim)
        self.lin2 = nn.Linear(mlp_dim, embedding_dim)
        self.act = act()

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        return self.lin2(self.act(self.lin1(x)))

class LayerNorm2d(nn.Module):
    def __init__(self, num_channels: int, eps: float = 1e-6) -> None:
        super().__init__()
        self.weight = nn.Parameter(torch.ones(num_channels))
        self.bias = nn.Parameter(torch.zeros(num_channels))
        self.eps = eps

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        u = x.mean(1, keepdim=True)
        s = (x - u).pow(2).mean(1, keepdim=True)
        x = (x - u) / torch.sqrt(s + self.eps)
        x = self.weight[:, None, None] * x + self.bias[:, None, None]
        return x





# This class and its supporting functions below lightly adapted from the ViTDet backbone available at: https://github.com/facebookresearch/detectron2/blob/main/detectron2/modeling/backbone/vit.py # noqa
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,
        include_neck: bool = False,
        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,
        )

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

        self.blocks = nn.ModuleList()
        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.blocks.append(block)
            
#         if self.include_neck :
        self.neck = nn.Sequential(
            nn.Conv2d(
                embed_dim,
                out_chans,
                kernel_size=1,
                bias=False,
            ),
            LayerNorm2d(out_chans),
            nn.Conv2d(
                out_chans,
                out_chans,
                kernel_size=3,
                padding=1,
                bias=False,
            ),
            LayerNorm2d(out_chans),
        )

    def forward(self, x: torch.Tensor) -> torch.Tensor:
#         x = self.patch_embed(x)
#         if self.pos_embed is not None:
#             x = x + self.pos_embed

        for blk in self.blocks:
            x = blk(x)
#             print(x.grad_fn)
        
#         if self.include_neck :
#             x = self.neck(x.permute(0, 3, 1, 2))
#         else :
#             x = x.permute(0, 3, 1, 2)
            
#         x = self.neck(x.permute(0, 3, 1, 2))
        x = x.permute(0, 3, 1, 2)
        return 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 (int or None): Input resolution for calculating the relative positional
                parameter size.
        """
        super().__init__()
        self.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.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 forward(self, x: torch.Tensor) -> torch.Tensor:
        shortcut = x
        x = self.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.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 (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 = nn.Parameter(torch.zeros(2 * input_size[0] - 1, head_dim))
            self.rel_pos_w = nn.Parameter(torch.zeros(2 * input_size[1] - 1, head_dim))

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        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).permute(2, 0, 3, 1, 4)
        # q, k, v with shape (B * nHead, H * W, C)
        q, k, v = qkv.reshape(3, B * self.num_heads, H * W, -1).unbind(0)

        attn = (q * self.scale) @ k.transpose(-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 = attn.softmax(dim=-1)
        x = (attn @ v).view(B, self.num_heads, H, W, -1).permute(0, 2, 3, 1, 4).reshape(B, H, W, -1)
        x = self.proj(x)

        return x


def window_partition(x: torch.Tensor, window_size: int) -> Tuple[torch.Tensor, Tuple[int, int]]:
    """
    Partition into non-overlapping windows with padding if needed.
    Args:
        x (tensor): 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 = F.pad(x, (0, 0, 0, pad_w, 0, pad_h))
    Hp, Wp = H + pad_h, W + pad_w

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


def window_unpartition(
    windows: torch.Tensor, window_size: int, pad_hw: Tuple[int, int], hw: Tuple[int, int]
) -> torch.Tensor:
    """
    Window unpartition into original sequences and removing padding.
    Args:
        x (tensor): 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.view(B, Hp // window_size, Wp // window_size, window_size, window_size, -1)
    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, Hp, Wp, -1)

    if Hp > H or Wp > W:
        x = x[:, :H, :W, :].contiguous()
    return x


def get_rel_pos(q_size: int, k_size: int, rel_pos: torch.Tensor) -> torch.Tensor:
#     """
#     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 (Tensor): 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 = F.interpolate(
            rel_pos.reshape(1, rel_pos.shape[0], -1).permute(0, 2, 1),
            size=max_rel_dist,
            mode="linear",
        )
        rel_pos_resized = rel_pos_resized.reshape(-1, max_rel_dist).permute(1, 0)
    else:
        rel_pos_resized = rel_pos

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


def add_decomposed_rel_pos(
    attn: torch.Tensor,
    q: torch.Tensor,
    rel_pos_h: torch.Tensor,
    rel_pos_w: torch.Tensor,
    q_size: Tuple[int, int],
    k_size: Tuple[int, int],
) -> torch.Tensor:
    """
    Calculate decomposed Relative Positional Embeddings from :paper:`mvitv2`.
    https://github.com/facebookresearch/mvit/blob/19786631e330df9f3622e5402b4a419a263a2c80/mvit/models/attention.py   # noqa B950
    Args:
        attn (Tensor): attention map.
        q (Tensor): query q in the attention layer with shape (B, q_h * q_w, C).
        rel_pos_h (Tensor): relative position embeddings (Lh, C) for height axis.
        rel_pos_w (Tensor): 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 (Tensor): 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)
    rel_h = torch.einsum("bhwc,hkc->bhwk", r_q, Rh)
    rel_w = torch.einsum("bhwc,wkc->bhwk", r_q, Rw)

    attn = (
        attn.view(B, q_h, q_w, k_h, k_w) + rel_h[:, :, :, :, None] + rel_w[:, :, :, None, :]
    ).view(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):  embed_dim (int): Patch embedding dimension.
        """
        super().__init__()

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

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        x = self.proj(x)
        # B C H W -> B H W C
        x = x.permute(0, 2, 3, 1)
        return x
    
    
    
if __name__ == '__main__':
    import torch
    transunet = TransUNet(
                      in_channels=1,
                      out_channels=96,
                      pretrained ='ssl',
                      patch_dim = 16,
                      class_num=8)

    print(sum(p.numel() for p in transunet.parameters()))
    print(transunet(torch.randn(4, 1, 224, 224)).shape)