Python LSTM的输入维度

我有这样的模型

它从最后5个条目预测下一个条目

（这意味着给出

项目[0]

~

项目[4]

预测

项目[4]

）

我有培训数据

（9411026）

因此，我将

（93，511026）

作为输入，并将

（93，11026）

作为验证。 （然后将测试和数值分开）

然后 历史=model.fit（x，y，epochs=epoch，批量大小=10，验证数据=（valux，valuy））

然而，这种错误发生了

我猜它是说我应该先给出形状

[1011026]

，而不是

[10,511026]

然而，我想做的是给他们最后5个，并预测下一个

为什么错了？（我认为它适用于SimpleRN）

模型：“顺序”
_________________________________________________________________
层（类型）输出形状参数
=================================================================
lstm（lstm）（无，5512）23631872
_________________________________________________________________
致密的（致密的）（无，5512）262656
_________________________________________________________________
稠密的（稠密的）（无，511026）5656338
=================================================================
总参数：29550866
可培训参数：29550866
不可训练参数：0
_________________________________________________________________
回溯（最近一次呼叫最后一次）：
文件“manage.py”，第22行，在
main（）
文件“manage.py”，第18行，主
从命令行（sys.argv）执行命令
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/_-init____;.py”，第401行，从命令行执行
utility.execute（）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/_-init___;.py”，第395行，在execute中
self.fetch_命令（子命令）。从_argv（self.argv）运行_
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/base.py”，第330行，运行时来自
self.execute（*args，**cmd_选项）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/django/core/management/base.py”，第371行，在execute中
输出=self.handle（*args，**选项）
文件“/Users/whitebear/CodingWorks/httproot/aiwave/defapp/management/commands/learn.py”，第205行，在handle中
历史=model.fit（f.x，f.y，epochs=epoch，批量大小=10，验证数据=（f.val\u x，f.val\u y））
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/keras/engine/training.py”，第108行，在方法包装中
返回方法（self、*args、**kwargs）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/keras/engine/training.py”，第1098行
tmp_logs=训练函数（迭代器）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/def_function.py”，第780行，in_u调用__
结果=自身调用（*args，**kwds）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/def_function.py”，第840行，in_call
返回self.\u无状态\u fn（*args，**kwds）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/function.py”，第2829行，在调用中__
返回图形\函数。\过滤\调用（args，kwargs）\ pylint:disable=受保护的访问
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/function.py”，第1848行，在“过滤”调用中
取消管理器=取消管理器）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/function.py”，第1924行，在调用平面中
ctx，args，取消管理器=取消管理器）
调用中的第550行文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/function.py”
ctx=ctx）
文件“/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site packages/tensorflow/python/eager/execute.py”，第60行，快速执行
输入、属性、数量（输出）
tensorflow.python.framework.errors\u impl.InvalidArgumentError:不兼容的形状：[10,511026]与[1011026]
[[node gradient_tape/mean_squared_error/BroadcastGradientArgs（定义于/Users/whitebear/CodingWorks/httproot/aiwave/defapp/management/commands/learn.py:205）][Op:[Op:[Uu推断\训练\函数]
函数调用堆栈：
列车功能

您的模型预测的是5项，而不是1项

尝试：

---

哇，谢谢你，那是我的粗心。我删除

return\u sequences=True

，它就可以工作了

    n_hidden = 512
    model = Sequential()
    model.add(LSTM(n_hidden, activation=None, input_shape=(5,11026), return_sequences=True))
    model.add(Dense(n_hidden, activation="linear")) 
    model.add(Dense(n_in, activation="linear"))
    opt = Adam(lr=0.001)
    model.compile(loss='mse', optimizer=opt)
    model.summary()

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
lstm (LSTM)                  (None, 5, 512)            23631872  
_________________________________________________________________
dense (Dense)                (None, 5, 512)            262656    
_________________________________________________________________
dense_1 (Dense)              (None, 5, 11026)          5656338   
=================================================================
Total params: 29,550,866
Trainable params: 29,550,866
Non-trainable params: 0
_________________________________________________________________
Traceback (most recent call last):
  File "manage.py", line 22, in <module>
    main()
  File "manage.py", line 18, in main
    execute_from_command_line(sys.argv)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/__init__.py", line 401, in execute_from_command_line
    utility.execute()
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/__init__.py", line 395, in execute
    self.fetch_command(subcommand).run_from_argv(self.argv)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/base.py", line 330, in run_from_argv
    self.execute(*args, **cmd_options)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/django/core/management/base.py", line 371, in execute
    output = self.handle(*args, **options)
  File "/Users/whitebear/CodingWorks/httproot/aiwave/defapp/management/commands/learn.py", line 205, in handle
    history = model.fit(f.x, f.y, epochs=epoch, batch_size=10,validation_data=(f.val_x, f.val_y))
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/keras/engine/training.py", line 108, in _method_wrapper
    return method(self, *args, **kwargs)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/keras/engine/training.py", line 1098, in fit
    tmp_logs = train_function(iterator)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/def_function.py", line 780, in __call__
    result = self._call(*args, **kwds)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/def_function.py", line 840, in _call
    return self._stateless_fn(*args, **kwds)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/function.py", line 2829, in __call__
    return graph_function._filtered_call(args, kwargs)  # pylint: disable=protected-access
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/function.py", line 1848, in _filtered_call
    cancellation_manager=cancellation_manager)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/function.py", line 1924, in _call_flat
    ctx, args, cancellation_manager=cancellation_manager))
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/function.py", line 550, in call
    ctx=ctx)
  File "/Users/whitebear/anaconda3/envs/aiwave/lib/python3.6/site-packages/tensorflow/python/eager/execute.py", line 60, in quick_execute
    inputs, attrs, num_outputs)
tensorflow.python.framework.errors_impl.InvalidArgumentError:  Incompatible shapes: [10,5,11026] vs. [10,11026]
     [[node gradient_tape/mean_squared_error/BroadcastGradientArgs (defined at /Users/whitebear/CodingWorks/httproot/aiwave/defapp/management/commands/learn.py:205) ]] [Op:__inference_train_function_2084]

Function call stack:
train_function

model.add(LSTM(n_hidden, activation=None, input_shape=(5,11026)))