发生了异常：InvalidArgumentError您必须使用dtype字符串和形状[1]

如何解决发生了异常：InvalidArgumentError您必须使用dtype字符串和形状[1]

我只写了我的代码中出现占位符的错误。我在下面的错误代码前加注

# Read the shape information and compressed string from the binary file. 
string_ll= tf.placeholder(tf.string,[1])
string_lh= tf.placeholder(tf.string,[1])
string_hl= tf.placeholder(tf.string,[1])
string_hh= tf.placeholder(tf.string,[1])

x_ll_shape = tf.placeholder(tf.int32,[2])
x_lh_shape = tf.placeholder(tf.int32,[2])
x_hl_shape = tf.placeholder(tf.int32,[2])
x_hh_shape = tf.placeholder(tf.int32,[2])

y_ll_shape = tf.placeholder(tf.int32,[2])
y_lh_shape = tf.placeholder(tf.int32,[2])
y_hl_shape = tf.placeholder(tf.int32,[2])
y_hh_shape = tf.placeholder(tf.int32,[2])

with open((args.input_file+'ll'),"rb") as f:
 packed_ll = tfc.PackedTensors(f.read())
with open((args.input_file+'lh'),"rb") as f:
 packed_lh = tfc.PackedTensors(f.read())
with open((args.input_file+'hl'),"rb") as f:
 packed_hl = tfc.PackedTensors(f.read())
with open((args.input_file+'hh'),"rb") as f:
packed_hh = tfc.PackedTensors(f.read()) 

tensors_ll = [string_ll,x_ll_shape,y_ll_shape]
tensors_lh = [string_lh,x_lh_shape,y_lh_shape]
tensors_hl = [string_hl,x_hl_shape,y_hl_shape]
tensors_hh = [string_hh,x_hh_shape,y_hh_shape]

arrays_ll = packed_ll.unpack(tensors_ll)
arrays_lh = packed_lh.unpack(tensors_lh)
arrays_hl = packed_hl.unpack(tensors_hl)
arrays_hh = packed_hh.unpack(tensors_hh)

# Instantiate model.
prediction=Prediction(args.num_filters)
encoder = Encoder(args.num_filters)
decoder = Decoder(args.num_filters)
entropy_bottleneck = tfc.EntropyBottleneck(dtype=tf.float32)

y_ll_shape=tf.concat([y_ll_shape,[args.num_filters]],axis=0)
y_lh_shape=tf.concat([y_lh_shape,axis=0)
y_hl_shape=tf.concat([y_hl_shape,axis=0)
y_hh_shape=tf.concat([y_hh_shape,axis=0)

y_ll_hat = entropy_bottleneck.decompress(string_ll,y_ll_shape,channels=args.num_filters)
y_lh_hat = entropy_bottleneck.decompress(string_lh,y_lh_shape,channels=args.num_filters)
y_hl_hat = entropy_bottleneck.decompress(string_hl,y_hl_shape,channels=args.num_filters)
y_hh_hat = entropy_bottleneck.decompress(string_hh,y_hh_shape,channels=args.num_filters)

x_ll_hat = decoder(y_ll_hat)
x_lh_hat = decoder(y_lh_hat)
x_hl_hat = decoder(y_hl_hat)
x_hh_hat = decoder(y_hh_hat)

p_ff_hat=prediction(x_ll_hat)

x_lh_hat= x_lh_hat+p_ff_hat #here when i do the operation without add of p_ff_hat variable      
x_hl_hat= x_hl_hat+p_ff_hat #it look fine but in addition of the variable these line of code 
x_hh_hat= x_hh_hat+p_ff_hat  #execute but the error i got in last session. i mention 

# Remove batch dimension,and crop away any extraneous padding on the bottom
x_ll_hat = x_ll_hat[0,:x_ll_shape[0],:x_ll_shape[1],:]
x_lh_hat = x_lh_hat[0,:x_lh_shape[0],:x_lh_shape[1],:]
x_hl_hat = x_hl_hat[0,:x_hl_shape[0],:x_hl_shape[1],:]
x_hh_hat = x_hh_hat[0,:x_hh_shape[0],:x_hh_shape[1],:]
 

op_ll = write_png(args.output_file+'ll',x_ll_hat)
op_lh = write_png(args.output_file+'lh',x_lh_hat)
op_hl = write_png(args.output_file+'hl',x_hl_hat)
op_hh = write_png(args.output_file+'hh',x_hh_hat)

# Load the latest model checkpoint,and perform the above actions.
with tf.Session() as sess:
   latest = tf.train.latest_checkpoint(checkpoint_dir=args.checkpoint_dir)
   tf.train.Saver().restore(sess,save_path=latest)

   sess.run(op_ll,feed_dict=dict(zip(tensors_ll,arrays_ll)))

   sess.run(op_lh,feed_dict=dict(zip(tensors_lh,arrays_lh))) #here I got the error
   sess.run(op_hl,feed_dict=dict(zip(tensors_hl,arrays_hl)))
   sess.run(op_hh,feed_dict=dict(zip(tensors_hh,arrays_hh)))

我得到的错误如下：
我被困在这里，因为从两周后起，您的帮助将受到真正的感谢，这对我来说非常重要。谢谢

找到2个根错误。
（0）无效的参数：您必须使用dtype字符串和形状[1] [[node占位符（在home / image902 / Desktop / compression-master / models / wavelet.py：783中定义）”为占位符张量“占位符”输入值]]
（1）无效的参数：您必须使用dtype字符串和形状[1] [[节点占位符（在home / image902 / Desktop / compression-master / models / wavelet.py：783中定义）”为占位符张量“占位符”输入值]] [[entropy_bottleneck_1 / decompress / TensorArray_1 / _217]]
0次成功操作。
忽略0个派生错误。

“占位符”的原始堆栈跟踪：
_run_module_as_main“ main ”，mod_spec中的文件“ usr / lib / python3.6 / runpy.py”，第193行，
_run_code
中的文件“ usr / lib / python3.6 / runpy.py”，第85行 exec（代码，run_globals）
在文件
中的第45行中添加文件“ home / image902 / .vscode / extensions / ms-python.python 2020.9.114305 / pythonFiles / lib / python / debugpy / main .py” cli.main（）
在主文件中的文件“ home / image902 / .vscode / extensions / ms-python.python-2020.9.114305 / pythonFiles / lib / python / debugpy /../ debugpy / server / cli.py”中
run（）
在run_file
的第267行中，文件“ home / image902 / .vscode / extensions / ms-python.python-2020.9.114305 / pythonFiles / lib / python / debugpy /../ debugpy / server / cli.py” runpy.run_path（options.target，run_name = compat.force_str（“ 主要”））
在run_path pkg_name = pkg_name，script_name = fname中的文件“ usr / lib / python3.6 / runpy.py”，第263行，
在_run_module_code mod_name，mod_spec，pkg_name，script_name中的文件“ usr / lib / python3.6 / runpy.py”，第96行，
_run_code
中的文件“ usr / lib / python3.6 / runpy.py”，第85行 exec（代码，run_globals）

中的文件“ home / image902 / Desktop / compression-master / models / wavelet.py”，第1046行 main（args）
主文件991行中的文件“ home / image902 / Desktop / compression-master / models / wavelet.py”
解压缩（args）
解压缩文件“ home / image902 / Desktop / compression-master / models / wavelet.py”，第783行