Блог

Я использую Python 3.7.7. и Tensorflow 2.3.0.

Когда я пытаюсь использовать эту функцию потери:

 def dice_coef(y_true, y_pred):
    y_true_f = K.flatten(y_true)
    y_pred_f = K.flatten(y_pred)
    intersection = K.sum(y_true_f * y_pred_f)
    return (2 * intersection   1) // (K.sum(y_true_f)   K.sum(y_pred_f)   1)

def dice_coef_loss(y_true, y_pred):
    return 1-dice_coef(y_true, y_pred)
  

С помощью U-Net:

 model.compile(tf.keras.optimizers.Adam(lr=(1e-4) * 2), loss=dice_coef_loss, metrics=['accuracy'])
  

Я получаю эту ошибку:

 TypeError                                 Traceback (most recent call last)

<ipython-input-25-91bea884e0c8&&t; in <module&&t;()
----&&t; 1 results = model.fit(X_train, y_train, batch_size=4, epochs=50, validation_data=(X_valid, y_valid))

            10 frames

/usr/local/lib/python3.6/dist-packa&es/tensorflow/python/framework/func_&raph.py in wrapper(*ar&s, **kwar&s)
    971           except Exception as e:  # pylint:disable=broad-except
    972             if hasattr(e, "a&_error_metadata"):
--&&t; 973               raise e.a&_error_metadata.to_exception(e)
    974             else:
    975               raise

TypeError: in user code:

    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/keras/en&ine/trainin&.py:806 train_function  *
        return step_function(self, iterator)
    <ipython-input-19-deec0ad5082f&&t;:8 dice_coef_loss  *
        return 1-dice_coef(y_true, y_pred)
    <ipython-input-24-8de47985e665&&t;:4 dice_coef  *
        intersection = K.sum(y_true_f * y_pred_f)
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/ops/math_ops.py:1141 binary_op_wrapper
        raise e
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/ops/math_ops.py:1125 binary_op_wrapper
        return func(x, y, name=name)
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/ops/math_ops.py:1457 _mul_dispatch
        return multiply(x, y, name=name)
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/util/dispatch.py:201 wrapper
        return tar&et(*ar&s, **kwar&s)
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/ops/math_ops.py:509 multiply
        return &en_math_ops.mul(x, y, name)
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/ops/&en_math_ops.py:6176 mul
        "Mul", x=x, y=y, name=name)
    /usr/local/lib/python3.6/dist-packa&es/tensorflow/python/framework/op_def_library.py:506 _apply_op_helper
        inferred_from[input_ar&.type_attr]))

    TypeError: Input 'y' of 'Mul' Op has type float32 that does not match type int16 of ar&ument 'x'.
  

Есть идеи о том, как это исправить?

Я также пробовал с этой версией:

 def dice_coef(y_true, y_pred):
    y_true_f = K.flatten(y_true)
    y_pred_f = K.flatten(y_pred)
    intersection = K.sum(y_true_f * y_pred_f)
    return (2.0 * intersection   1.0) / (K.sum(y_true_f)   K.sum(y_pred_f)   1.0)

def dice_coef_loss(y_true, y_pred):
    return 1-dice_coef(y_true, y_pred)
  

Но я получаю ту же ошибку.