TensorFlow生成器适配112×112输入时Concatenate层报错问题

你使用的U-Net结构默认适配输入尺寸为2的整数次幂（如256=2^8）：8次步长为2的下采样可将256x256输入降至1x1，搭配后续7次上采样和最后1层转置卷积，刚好能还原为256x256输出。
将输入调整为112x112后，由于112不是2的整数次幂，多次下采样后的特征尺寸和上采样阶段的特征尺寸无法匹配，导致Concatenate层拼接时出现长宽维度的冲突，你收到的报错就是第一次上采样输出2x2特征时，对应的跳连特征尺寸为1x1，无法直接拼接。

提供两种可直接落地的修改方案：

方案1：调整网络层数量适配112x112输入 #

你使用的downsample（步长2的卷积）和upsample（步长2的转置卷积）默认每调用一次，特征长宽就会缩放一倍，112x112输入最多支持6次有效下采样，无需原有8层下采样结构。
修改点如下：

• 移除down_stack最后1层下采样，保留前7层
• 移除up_stack最后1层上采样，保留前6层
• 调整最后一层转置卷积的配置，确保输出尺寸符合你的需求
  修改后的核心代码参考：

def Generator():
  inputs = tf.keras.layers.Input(shape=[112, 112, 3])

  down_stack = [
    downsample(64, 4, apply_batchnorm=False),  # (batch_size, 56, 56, 64)
    downsample(128, 4),  # (batch_size, 28, 28, 128)
    downsample(256, 4),  # (batch_size, 14, 14, 256)
    downsample(512, 4),  # (batch_size, 7, 7, 512)
    downsample(512, 4),  # (batch_size, 4, 4, 512)
    downsample(512, 4),  # (batch_size, 2, 2, 512)
    downsample(512, 4),  # (batch_size, 1, 1, 512)
  ]

  up_stack = [
    upsample(512, 4, apply_dropout=True),  # (batch_size, 2, 2, 1024)
    upsample(512, 4, apply_dropout=True),  # (batch_size, 4, 4, 1024)
    upsample(512, 4, apply_dropout=True),  # (batch_size, 8, 8, 1024)
    upsample(512, 4),  # (batch_size, 16, 16, 1024)
    upsample(256, 4),  # (batch_size, 32, 32, 512)
    upsample(128, 4),  # (batch_size, 64, 64, 256)
  ]

  initializer = tf.random_normal_initializer(0., 0.02)
  last = tf.keras.layers.Conv2DTranspose(OUTPUT_CHANNELS, 4,
                                         strides=2,
                                         padding='same',
                                         kernel_initializer=initializer,
                                         activation='tanh')  # (batch_size, 112, 112, 3)

  x = inputs

  # Downsampling through the model
  skips = []
  for down in down_stack:
    x = down(x)
    skips.append(x)

  skips = reversed(skips[:-1])

  # Upsampling and establishing the skip connections
  for up, skip in zip(up_stack, skips):
    x = up(x)
    x = tf.keras.layers.Concatenate()([x, skip])

  x = last(x)

  return tf.keras.Model(inputs=inputs, outputs=x)

方案2：拼接前做尺寸对齐，无需调整网络层数 #

如果不想改动原有网络结构，直接在跳连拼接前增加尺寸对齐逻辑即可，改动量最小，适合快速验证效果：
仅需要修改跳连部分的代码：

# Upsampling and establishing the skip connections
for up, skip in zip(up_stack, skips):
    x = up(x)
    # 新增尺寸对齐逻辑
    if x.shape[1] != skip.shape[1] or x.shape[2] != skip.shape[2]:
        skip = tf.keras.layers.Resizing(height=x.shape[1], width=x.shape[2])(skip)
    x = tf.keras.layers.Concatenate()([x, skip])

内容的提问来源于stack exchange，提问作者Hazem