Spaces:
Paused
Paused
| import torch | |
| import torch.nn as nn | |
| import numpy as np | |
| from einops import rearrange, repeat | |
| class Residual(nn.Module): | |
| def __init__(self, fn): | |
| super().__init__() | |
| self.fn = fn | |
| def forward(self, x, **kwargs): | |
| return self.fn(x, **kwargs) + x | |
| class PreNorm(nn.Module): | |
| def __init__(self, dim, fn): | |
| super().__init__() | |
| self.norm = nn.LayerNorm(dim) | |
| self.fn = fn | |
| def forward(self, x, **kwargs): | |
| return self.fn(self.norm(x), **kwargs) | |
| class FeedForward(nn.Module): | |
| def __init__(self, dim, hidden_dim, dropout = 0.): | |
| super().__init__() | |
| self.net = nn.Sequential( | |
| nn.Linear(dim, hidden_dim), | |
| nn.GELU(), | |
| nn.Dropout(dropout), | |
| nn.Linear(hidden_dim, dim), | |
| nn.Dropout(dropout) | |
| ) | |
| def forward(self, x): | |
| return self.net(x) | |
| class Attention(nn.Module): | |
| def __init__(self, dim, heads, dim_head, dropout): | |
| super().__init__() | |
| inner_dim = dim_head * heads | |
| self.heads = heads | |
| self.scale = dim_head ** -0.5 | |
| self.to_qkv = nn.Linear(dim, inner_dim * 3, bias = False) | |
| self.to_out = nn.Sequential( | |
| nn.Linear(inner_dim, dim), | |
| nn.Dropout(dropout) | |
| ) | |
| def forward(self, x, mask = None): | |
| # x:[b,n,dim] | |
| b, n, _, h = *x.shape, self.heads | |
| # get qkv tuple:([b,n,head_num*head_dim],[...],[...]) | |
| qkv = self.to_qkv(x).chunk(3, dim = -1) | |
| # split q,k,v from [b,n,head_num*head_dim] -> [b,head_num,n,head_dim] | |
| q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b h n d', h = h), qkv) | |
| # transpose(k) * q / sqrt(head_dim) -> [b,head_num,n,n] | |
| dots = torch.einsum('bhid,bhjd->bhij', q, k) * self.scale | |
| mask_value = -torch.finfo(dots.dtype).max | |
| # mask value: -inf | |
| if mask is not None: | |
| mask = F.pad(mask.flatten(1), (1, 0), value = True) | |
| assert mask.shape[-1] == dots.shape[-1], 'mask has incorrect dimensions' | |
| mask = mask[:, None, :] * mask[:, :, None] | |
| dots.masked_fill_(~mask, mask_value) | |
| del mask | |
| # softmax normalization -> attention matrix | |
| attn = dots.softmax(dim=-1) | |
| # value * attention matrix -> output | |
| out = torch.einsum('bhij,bhjd->bhid', attn, v) | |
| # cat all output -> [b, n, head_num*head_dim] | |
| out = rearrange(out, 'b h n d -> b n (h d)') | |
| out = self.to_out(out) | |
| return out | |
| class CrossAttention(nn.Module): | |
| def __init__(self, dim, heads, dim_head, dropout): | |
| super().__init__() | |
| inner_dim = dim_head * heads | |
| project_out = not (heads == 1 and dim_head == dim) | |
| self.heads = heads | |
| self.scale = dim_head ** -0.5 | |
| self.to_k = nn.Linear(dim, inner_dim , bias=False) | |
| self.to_v = nn.Linear(dim, inner_dim , bias = False) | |
| self.to_q = nn.Linear(dim, inner_dim, bias = False) | |
| self.to_out = nn.Sequential( | |
| nn.Linear(inner_dim, dim), | |
| nn.Dropout(dropout) | |
| ) if project_out else nn.Identity() | |
| def forward(self, x_qkv): | |
| b, n, _, h = *x_qkv.shape, self.heads | |
| k = self.to_k(x_qkv) | |
| k = rearrange(k, 'b n (h d) -> b h n d', h = h) | |
| v = self.to_v(x_qkv) | |
| v = rearrange(v, 'b n (h d) -> b h n d', h = h) | |
| q = self.to_q(x_qkv[:, 0].unsqueeze(1)) | |
| q = rearrange(q, 'b n (h d) -> b h n d', h = h) | |
| dots = torch.einsum('b h i d, b h j d -> b h i j', q, k) * self.scale | |
| attn = dots.softmax(dim=-1) | |
| out = torch.einsum('b h i j, b h j d -> b h i d', attn, v) | |
| out = rearrange(out, 'b h n d -> b n (h d)') | |
| out = self.to_out(out) | |
| return out | |
| class Transformer(nn.Module): | |
| def __init__(self, dim, depth, heads, dim_head, mlp_head, dropout, num_channel): | |
| super().__init__() | |
| self.layers = nn.ModuleList([]) | |
| for _ in range(depth): | |
| self.layers.append(nn.ModuleList([ | |
| Residual(PreNorm(dim, Attention(dim, heads = heads, dim_head = dim_head, dropout = dropout))), | |
| Residual(PreNorm(dim, FeedForward(dim, mlp_head, dropout = dropout))) | |
| ])) | |
| self.skipcat = nn.ModuleList([]) | |
| for _ in range(depth-2): | |
| self.skipcat.append(nn.Conv2d(num_channel+1, num_channel+1, [1, 2], 1, 0)) | |
| def forward(self, x, mask = None): | |
| for attn, ff in self.layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| return x | |
| class SSTransformer(nn.Module): | |
| def __init__(self, dim, depth, heads, dim_head, mlp_head, b_dim, b_depth, b_heads, b_dim_head, b_mlp_head, num_patches, dropout): | |
| super().__init__() | |
| self.layers = nn.ModuleList([]) | |
| self.k_layers = nn.ModuleList([]) | |
| self.channels_to_embedding = nn.Linear(num_patches, b_dim) | |
| self.cls_token = nn.Parameter(torch.randn(1, 1, b_dim)) | |
| for _ in range(depth): | |
| self.layers.append(nn.ModuleList([ | |
| Residual(PreNorm(dim, Attention(dim, heads = heads, dim_head = dim_head, dropout = dropout))), | |
| Residual(PreNorm(dim, FeedForward(dim, mlp_head, dropout = dropout))) | |
| ])) | |
| for _ in range(b_depth): | |
| self.k_layers.append(nn.ModuleList([ | |
| Residual(PreNorm(b_dim, Attention(dim=b_dim, heads=b_heads, dim_head=b_dim_head, dropout = dropout))), | |
| Residual(PreNorm(b_dim, FeedForward(b_dim, b_mlp_head, dropout = dropout))) | |
| ])) | |
| def forward(self, x, mask = None): | |
| for attn, ff in self.layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| x = rearrange(x, 'b n d -> b d n') | |
| x = self.channels_to_embedding(x) | |
| b, d, n = x.shape | |
| cls_tokens = repeat(self.cls_token, '() n d -> b n d', b = b) | |
| x = torch.cat((cls_tokens, x), dim = 1) | |
| for attn, ff in self.k_layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| return x | |
| class SSTransformer_pyramid(nn.Module): | |
| def __init__(self, dim, depth, heads, dim_head, mlp_head, b_dim, b_depth, b_heads, b_dim_head, b_mlp_head, num_patches, dropout): | |
| super().__init__() | |
| self.layers = nn.ModuleList([]) | |
| self.k_layers = nn.ModuleList([]) | |
| self.channels_to_embedding = nn.Linear(num_patches, b_dim) | |
| self.cls_token = nn.Parameter(torch.randn(1, 1, b_dim)) | |
| for _ in range(depth): | |
| self.layers.append(nn.ModuleList([ | |
| Residual(PreNorm(dim, Attention(dim, heads = heads, dim_head = dim_head, dropout = dropout))), | |
| Residual(PreNorm(dim, FeedForward(dim, mlp_head, dropout = dropout))) | |
| ])) | |
| for _ in range(b_depth): | |
| self.k_layers.append(nn.ModuleList([ | |
| Residual(PreNorm(b_dim, Attention(dim=b_dim, heads=b_heads, dim_head=b_dim_head, dropout = dropout))), | |
| Residual(PreNorm(b_dim, FeedForward(b_dim, b_mlp_head, dropout = dropout))) | |
| ])) | |
| def forward(self, x, mask = None): | |
| for attn, ff in self.layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| out_feature = x | |
| x = rearrange(x, 'b n d -> b d n') | |
| x = self.channels_to_embedding(x) | |
| b, d, n = x.shape | |
| cls_tokens = repeat(self.cls_token, '() n d -> b n d', b = b) | |
| x = torch.cat((cls_tokens, x), dim = 1) | |
| for attn, ff in self.k_layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| return x, out_feature | |
| class ViT(nn.Module): | |
| def __init__(self, image_size, near_band, num_patches, num_classes, dim, depth, heads, mlp_dim, pool='cls', channel_dim=1, dim_head = 16, dropout=0., emb_dropout=0., mode='ViT'): | |
| super().__init__() | |
| patch_dim = image_size ** 2 * near_band | |
| self.pos_embedding = nn.Parameter(torch.randn(1, num_patches + 1, dim)) | |
| self.patch_to_embedding = nn.Linear(channel_dim, dim) | |
| self.cls_token = nn.Parameter(torch.randn(1, 1, dim)) | |
| self.dropout = nn.Dropout(emb_dropout) | |
| self.transformer = Transformer(dim, depth, heads, dim_head, mlp_dim, dropout, num_patches, mode) | |
| self.pool = pool | |
| self.to_latent = nn.Identity() | |
| self.mlp_head = nn.Sequential( | |
| nn.LayerNorm(dim), | |
| nn.Linear(dim, num_classes) | |
| ) | |
| def forward(self, x, mask = None): | |
| # patchs[batch, patch_num, patch_size*patch_size*c] [batch,200,145*145] | |
| # x = rearrange(x, 'b c h w -> b c (h w)') | |
| ## embedding every patch vector to embedding size: [batch, patch_num, embedding_size] | |
| x = self.patch_to_embedding(x) #[b,n,dim] | |
| b, n, _ = x.shape | |
| # add position embedding | |
| cls_tokens = repeat(self.cls_token, '() n d -> b n d', b = b) #[b,1,dim] | |
| x = torch.cat((cls_tokens, x), dim = 1) #[b,n+1,dim] | |
| x += self.pos_embedding[:, :(n + 1)] | |
| x = self.dropout(x) | |
| # transformer: x[b,n + 1,dim] -> x[b,n + 1,dim] | |
| x = self.transformer(x, mask) | |
| # classification: using cls_token output | |
| x = self.to_latent(x[:,0]) | |
| # MLP classification layer | |
| return self.mlp_head(x) | |
| class SSFormer_v4(nn.Module): | |
| def __init__(self, dim, depth, heads, dim_head, mlp_head, b_dim, b_depth, b_heads, b_dim_head, b_mlp_head, num_patches, dropout, mode): | |
| super().__init__() | |
| self.layers = nn.ModuleList([]) | |
| self.k_layers = nn.ModuleList([]) | |
| self.channels_to_embedding = nn.Linear(num_patches, b_dim) | |
| self.cls_token = nn.Parameter(torch.randn(1, 1, b_dim)) | |
| for _ in range(depth): | |
| self.layers.append(nn.ModuleList([ | |
| Residual(PreNorm(dim, Attention(dim, heads = heads, dim_head = dim_head, dropout = dropout))), | |
| Residual(PreNorm(dim, FeedForward(dim, mlp_head, dropout = dropout))) | |
| ])) | |
| for _ in range(b_depth): | |
| self.k_layers.append(nn.ModuleList([ | |
| Residual(PreNorm(b_dim, Attention(dim=b_dim, heads=b_heads, dim_head=b_dim_head, dropout = dropout))), | |
| Residual(PreNorm(b_dim, FeedForward(b_dim, b_mlp_head, dropout = dropout))) | |
| ])) | |
| self.mode = mode | |
| def forward(self, x, c, mask = None): | |
| for attn, ff in self.layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| x = rearrange(x, 'b n d -> b d n') | |
| x = self.channels_to_embedding(x) | |
| b, d, n = x.shape | |
| cls_tokens = repeat(c, '() n d -> b n d', b = b) | |
| x = torch.cat((cls_tokens, x), dim = 1) | |
| for attn, ff in self.k_layers: | |
| x = attn(x, mask = mask) | |
| x = ff(x) | |
| return x | |