ValueError for Sequential model.fit

Question

I’m learning Tensorflow and i tried out this program
This program is from this Youtube Video at 4:40:00

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt

x = np.array([float(i) for i in range(-7, 15, 3)])
y = x + 10
plt.scatter(x, y)
plt.show()

x = tf.constant(x)
y = tf.constant(y)

model = tf.keras.Sequential([
  tf.keras.layers.Dense(1)
])

model.compile(loss = tf.keras.losses.mae, 
              optimizer = tf.keras.optimizers.SGD(), 
              metrics = ["mae"])
model.fit(x, y, epochs = 5)

But I get this error:

Epoch 1/5
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-146-19ee806b894a> in <module>()
      6               optimizer = tf.keras.optimizers.SGD(),
      7               metrics = ["mae"])
----> 8 model.fit(x, y, epochs = 5)

1 frames
/usr/local/lib/python3.7/dist-packages/tensorflow/python/framework/func_graph.py in autograph_handler(*args, **kwargs)
   1145           except Exception as e:  # pylint_disable=broad-except
   1146             if hasattr(e, "ag_error_metadata"):
-> 1147               raise e.ag_error_metadata.to_exception(e)
   1148             else:
   1149               raise

ValueError: in user code:

    File "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py", line 1021, in train_function  *
        return step_function(self, iterator)
    File "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py", line 1010, in step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    File "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py", line 1000, in run_step  **
        outputs = model.train_step(data)
    File "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py", line 859, in train_step
        y_pred = self(x, training=True)
    File "/usr/local/lib/python3.7/dist-packages/keras/utils/traceback_utils.py", line 67, in error_handler
        raise e.with_traceback(filtered_tb) from None
    File "/usr/local/lib/python3.7/dist-packages/keras/engine/input_spec.py", line 228, in assert_input_compatibility
        raise ValueError(f'Input {input_index} of layer "{layer_name}" '

    ValueError: Exception encountered when calling layer "sequential_44" (type Sequential).
    
    Input 0 of layer "dense_39" is incompatible with the layer: expected min_ndim=2, found ndim=1. Full shape received: (None,)
    
    Call arguments received:
      • inputs=tf.Tensor(shape=(None,), dtype=float64)
      • training=True
      • mask=None

What mistake have i done and what should i do to correct it?
If I have to make some changes then how did it work in the video?

Asked By: Ruthvik

||

Source

Answer 1

You may need to add a new axis to x to be compatible with the dense layer. Try the following code.

model.compile(loss = tf.keras.losses.mae, 
              optimizer = tf.keras.optimizers.SGD(), 
              metrics = ["mae"])
model.fit(x[..., None], y, epochs = 100) 
# OK

Note, in the tutorials, the instructor also mentioned it in his Github repo.

Answered By: M.Innat

Answer 2

You need to add an input layer into your sequential model as follows:
model = tf.keras.Sequential([tf.keras.Input(shape=(1,)), tf.keras.layers.Dense(1) ]).

The input layer specifies that indeed you are expecting scalar inputs.

Answered By: Sam Ngugi

Answer 3

Thanks for posting this answer.

After applying your solution – I also learned that order is critical:

this order ERRORS out

model = tf.keras.Sequential()
model.add(tf.keras.layers.Dense(1))
model.add(tf.keras.Input(shape=(1,)))

Input shape has to come before Dense

model = tf.keras.Sequential()
model.add(tf.keras.Input(shape=(1,)))
model.add(tf.keras.layers.Dense(1))

— The answer model.fit(x[..., None], y, epochs = 100) fixes the problem too

Answered By: Macirish

ValueError for Sequential model.fit

Question:

Answers: