神经网络代码运行报错：输入形状不兼容问题排查

问题场景 #

运行以下代码训练二分类神经网络时，Epoch 1阶段触发ValueError，提示输入形状不兼容：期望形状为(None, 455, 30)，实际输入形状为(None, 30)。

原始代码 #

import pandas as pd
import tensorflow as tf
from sklearn.model_selection import train_test_split
dataset = pd.read_csv ("/content/dataset/cancer.csv")
x = dataset.drop(columns = ["diagnosis(1=m, 0=b)"])
y = dataset["diagnosis(1=m, 0=b)"]
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)
model = tf.keras.models.Sequential()
model.add(tf.keras.layers.Dense(256, input_shape = x_train.shape, activation="sigmoid"))
model.add(tf.keras.layers.Dense(256, activation="sigmoid"))
model.add(tf.keras.layers.Dense(1, activation="sigmoid"))
model.compile(optimizer="adam", loss="binary_crossentropy", metrics=["accuracy"])
model.fit(x_train, y_train, epochs=1000)

错误信息 #

Epoch 1/1000
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-45-a5e6a9bff808> in <cell line: 1>()
----> 1 model.fit(x_train, y_train, epochs=1000)

1 frames
/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py in tf__train_function(iterator)
     13                 try:
     14                     do_return = True
---&gt; 15                     retval_ = ag__.converted_call(ag__.ld(step_function), (ag__.ld(self), ag__.ld(iterator)), None, fscope)
     16                 except:
     17                     do_return = False

ValueError: in user code:

    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 1377, in train_function  *
        return step_function(self, iterator)
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 1360, in step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 1349, in run_step  **
        outputs = model.train_step(data)
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py", line 1126, in train_step
        y_pred = self(x, training=True)
    File "/usr/local/lib/python3.10/dist-packages/keras/src/utils/traceback_utils.py", line 70, in error_handler
        raise e.with_traceback(filtered_tb) from None
    File "/usr/local/lib/python3.10/dist-packages/keras/src/engine/input_spec.py", line 298, in assert_input_compatibility
        raise ValueError(

    ValueError: Input 0 of layer "sequential_10" is incompatible with the layer: expected shape=(None, 455, 30), found shape=(None, 30)

原因分析 #

错误出在model.add(tf.keras.layers.Dense(256, input_shape = x_train.shape, activation="sigmoid"))这一行：

• x_train.shape返回的是整个训练集的维度，即(455, 30)，其中455是训练样本数量，30是单样本的特征数。
• Keras的input_shape参数要求的是单个样本的特征维度，不需要包含样本数量。当你把整个训练集的shape传入后，模型错误地认为每个样本的形状是(455, 30)，但实际训练时输入的每个样本是一维的30个特征，形状为(30,)，导致输入形状不匹配。

修正方案 #

将input_shape = x_train.shape修改为input_shape = x_train.shape[1:]（自动获取单样本特征维度），或者直接写input_shape=(30,)。

修正后代码 #

import pandas as pd
import tensorflow as tf
from sklearn.model_selection import train_test_split
dataset = pd.read_csv ("/content/dataset/cancer.csv")
x = dataset.drop(columns = ["diagnosis(1=m, 0=b)"])
y = dataset["diagnosis(1=m, 0=b)"]
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2)
model = tf.keras.models.Sequential()
# 修改input_shape为单样本特征维度
model.add(tf.keras.layers.Dense(256, input_shape = x_train.shape[1:], activation="sigmoid"))
model.add(tf.keras.layers.Dense(256, activation="sigmoid"))
model.add(tf.keras.layers.Dense(1, activation="sigmoid"))
model.compile(optimizer="adam", loss="binary_crossentropy", metrics=["accuracy"])
model.fit(x_train, y_train, epochs=1000)

内容的提问来源于stack exchange，提问作者Adeep Sudarsan