“MergeSummary”Op的“inputs”具有与预期类型字符串不匹配的类型[bool，bool，bool，bool，bool，bool，bool]

Question

我得到这个错误：

ValueError: Tensor conversion requested dtype string for Tensor with dtype bool: <tf.Tensor 'lr/write_summary/Const:0' shape=() dtype=bool>

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/user/iibi/skuanar/Downloads/VAE-GAN-Autoencoding-Beyond-Pixels-Using-a-Similarity-Metric-master/train_vae_gan.py", line 30, in <module>
    training = train_stuff(gan)
  File "/nfs/s-iibi54/users/skuanar/Downloads/VAE-GAN-Autoencoding-Beyond-Pixels-Using-a-Similarity-Metric-master/vaegan.py", line 245, in train_stuff
    summ_op = tf.compat.v1.summary.merge(summaries)
  File "/user/iibi/skuanar/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/summary/summary.py", line 371, in merge
    val = _gen_logging_ops.merge_summary(inputs=inputs, name=name)
  File "/user/iibi/skuanar/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/ops/gen_logging_ops.py", line 530, in merge_summary
    "MergeSummary", inputs=inputs, name=name)
  File "/user/iibi/skuanar/anaconda3/lib/python3.7/site-packages/tensorflow_core/python/framework/op_def_library.py", line 493, in _apply_op_helper
    (prefix, dtype.name))
TypeError: Tensors in list passed to 'inputs' of 'MergeSummary' Op have types [bool, bool, bool, bool, bool, bool, bool, bool, bool] that do not match expected type string.

我的代码：

def train_stuff(model):
    #global_step = tf.train.create_global_step()
    global_step = tf.compat.v1.train.create_global_step()

    # learning_rate = tf.train.exponential_decay(config.learning_rate, global_step=global_step, decay_steps=10000, decay_rate=0.98)
    learning_rate = tf.compat.v1.train.exponential_decay(config.learning_rate, global_step=global_step, decay_steps=10000, decay_rate=0.98)

    train_vars = tf.compat.v1.trainable_variables()
    dec_vars = [var for var in train_vars if "decoder" in var.name]
    dis_vars = [var for var in train_vars if "discriminator" in var.name]
    enc_vars = [var for var in train_vars if "encoder" in var.name]
    dec_opt = tf.compat.v1.train.RMSPropOptimizer(learning_rate=learning_rate)
    dis_opt = tf.compat.v1.train.RMSPropOptimizer(learning_rate=learning_rate)
    enc_opt = tf.compat.v1.train.RMSPropOptimizer(learning_rate=learning_rate)
    dec_grad = dec_opt.compute_gradients(model["dec_loss"], var_list=dec_vars)
    dis_grad = dis_opt.compute_gradients(model["dis_loss"], var_list=dis_vars)
    enc_grad = enc_opt.compute_gradients(model["enc_loss"], var_list=enc_vars)
    dec_op = dec_opt.apply_gradients(dec_grad)
    dis_op = dis_opt.apply_gradients(dis_grad)
    enc_op = enc_opt.apply_gradients(enc_grad, global_step=global_step)

    summaries = [tf.summary.scalar("lr", learning_rate)] + model["summaries"]
    summ_op = tf.compat.v1.summary.merge(summaries)
    init = tf.compat.v1.global_variables_initializer()
    saver = tf.compat.v1.train.Saver()

    stuff = {"global_step": global_step, "dec_op": dec_op, "dis_op": dis_op, "summ_op": summ_op, "init": init, "saver": saver, "enc_op": enc_op}

    return stuff