Python 如何解决dtype float64的张量转换所需的dtype float32

我正在尝试重命名模型，并将其放在“bnn”中，以便以后使用 这是我的错误消息：

Tensor conversion requested dtype float32 for Tensor with dtype float64: <tf.Tensor 'dense_variational_4/sequential/multivariate_normal_tri_l/dense_variational_4_sequential_multivariate_normal_tri_l_MultivariateNormalTriL_MultivariateNormalTriL/value/dense_variational_4_sequential_multivariate_normal_tri_l_MultivariateNormalTriL_MultivariateNormalTriL/sample/dense_variational_4_sequential_multivariate_normal_tri_l_MultivariateNormalTriL_MultivariateNormalTriL_chain_of_dense_variational_4_sequential_multivariate_normal_tri_l_MultivariateNormalTriL_MultivariateNormalTriL_shift_of_dense_variational_4_sequential_multivariate_normal_tri_l_MultivariateNormalTriL_MultivariateNormalTriL_scale_matvec_linear_operator/forward/dense_variational_4_sequential_multivariate_normal_tri_l_MultivariateNormalTriL_MultivariateNormalTriL_shift/forward/add:0' shape=(120,) dtype=float64>

def create_bnn_model(train_size):
X_input, X_input_ = get_batch2(stock_path, batch_size)
inputs = create_model_inputs(X_input, X_input_)
features = keras.layers.concatenate(list(inputs.values()))
#features = layers.BatchNormalization()(features)



for layers in hidden_layers:
    features = tfp.layers.DenseVariational(
        units=layers,
        make_prior_fn=prior,
        make_posterior_fn=posterior,
        kl_weight= 1 / train_size,
        activation= tf.tanh,
    )(features)
    
# The output is deterministic: a single point estimate.
outputs = layers.Dense(units=1)(features)
model = keras.Model(inputs=inputs, outputs=outputs)
return model
bnn = create_bnn_model(train_size)