| 100 | |
| 101 | |
| 102 | class EncoderBlock(nn.Module): |
| 103 | def __init__(self, in_dim, mlp_dim, num_heads, dropout_rate=0.1, attn_dropout_rate=0.1): |
| 104 | super(EncoderBlock, self).__init__() |
| 105 | |
| 106 | self.norm1 = nn.LayerNorm(in_dim) |
| 107 | self.attn = SelfAttention(in_dim, heads=num_heads, dropout_rate=attn_dropout_rate) |
| 108 | if dropout_rate > 0: |
| 109 | self.dropout = nn.Dropout(dropout_rate) |
| 110 | else: |
| 111 | self.dropout = None |
| 112 | self.norm2 = nn.LayerNorm(in_dim) |
| 113 | self.mlp = MlpBlock(in_dim, mlp_dim, in_dim, dropout_rate) |
| 114 | |
| 115 | def forward(self, x): |
| 116 | residual = x |
| 117 | out = self.norm1(x) |
| 118 | out = self.attn(out) |
| 119 | if self.dropout: |
| 120 | out = self.dropout(out) |
| 121 | out += residual |
| 122 | residual = out |
| 123 | |
| 124 | out = self.norm2(out) |
| 125 | out = self.mlp(out) |
| 126 | out += residual |
| 127 | return out |
| 128 | |
| 129 | |
| 130 | class Encoder(nn.Module): |