Tensorflow Keras モデルの訓練でのエラー

Question

前提・実現したいこと

回帰：燃費を予測する | TensorFlow Core
この記事を参考にして自分なりにKerasを用いてみようと思ったのですが、以下のエラーが出てしまいました。

発生している問題・エラーメッセージ

2021-09-17 20:10:31.091616: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)
Traceback (most recent call last):
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\main.py", line 83, in <module>
    history = model.fit(
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\training.py", line 1184, in fit
    tmp_logs = self.train_function(iterator)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\def_function.py", line 885, in __call__
    result = self._call(*args, **kwds)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\def_function.py", line 933, in _call
    self._initialize(args, kwds, add_initializers_to=initializers)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\def_function.py", line 759, in _initialize
    self._stateful_fn._get_concrete_function_internal_garbage_collected(  # pylint: disable=protected-access
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\function.py", line 3066, in _get_concrete_function_internal_garbage_collected
    graph_function, _ = self._maybe_define_function(args, kwargs)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\function.py", line 3463, in _maybe_define_function
    graph_function = self._create_graph_function(args, kwargs)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\function.py", line 3298, in _create_graph_function
    func_graph_module.func_graph_from_py_func(
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\framework\func_graph.py", line 1007, in func_graph_from_py_func
    func_outputs = python_func(*func_args, **func_kwargs)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\eager\def_function.py", line 668, in wrapped_fn
    out = weak_wrapped_fn().__wrapped__(*args, **kwds)
  File "C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\framework\func_graph.py", line 994, in wrapper
    raise e.ag_error_metadata.to_exception(e)
ValueError: in user code:

    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\training.py:853 train_function  *
        return step_function(self, iterator)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\training.py:842 step_function  **
        outputs = model.distribute_strategy.run(run_step, args=(data,))
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:1286 run
        return self._extended.call_for_each_replica(fn, args=args, kwargs=kwargs)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:2849 call_for_each_replica
        return self._call_for_each_replica(fn, args, kwargs)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\tensorflow\python\distribute\distribute_lib.py:3632 _call_for_each_replica
        return fn(*args, **kwargs)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\training.py:835 run_step  **
        outputs = model.train_step(data)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\training.py:787 train_step
        y_pred = self(x, training=True)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\base_layer.py:1020 __call__
        input_spec.assert_input_compatibility(self.input_spec, inputs, self.name)
    C:\Users\IshitobiHyo\PycharmProjects\WeatherAINeu\venv\lib\site-packages\keras\engine\input_spec.py:250 assert_input_compatibility
        raise ValueError(

    ValueError: Input 0 of layer sequential is incompatible with the layer: expected axis -1 of input shape to have value 10 but received input with shape (None, 11)

該当のソースコード

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
from tensorflow import keras
from tensorflow.keras.callbacks import Callback
from tensorflow.keras import Sequential
from tensorflow.keras import layers
from tensorflow.keras.optimizers import RMSprop


def create_data(file):
    column_names = [
        'ave_temp',
        'precipitation',
        'daylight',
        'wind',
        'ave_cloud_cover',
        'ave_press',
        'ave_humid',
        'ave_sea_press',
        'ave_local_press',
        'max_temp',
        'min_temp'
    ]
    data_ori = pd.read_csv(file, header=None)
    data_ori = data_ori.drop([0, 1, 2, 3, 15], axis=1)
    data_ori.columns = column_names

    return data_ori


csv = 'Nagoya.csv'
raw_dataset = create_data(csv)
raw_dataset.dropna()
dataset = raw_dataset.copy()
train_dataset = dataset.sample(frac=0.8, random_state=0)
test_dataset = dataset.drop(train_dataset.index)

train_stats = train_dataset.describe()
train_stats = train_stats.transpose()

label = 'ave_temp'
train_labels = train_dataset.pop(label)
test_labels = test_dataset.pop(label)


def norm(x):
    return (x - train_stats['mean']) / train_stats['std']


normed_train_data = norm(train_dataset)
normed_test_data = norm(test_dataset)


def build_model():
    model = Sequential([
        layers.Dense(64, activation='relu', input_shape=[len(train_dataset.keys())]),
        layers.Dense(64, activation='relu'),
        layers.Dense(1)
    ])
    optimizer = RMSprop(0.001)
    model.compile(loss='mse', optimizer=optimizer, metrics=['mae', 'mse'])
    return model


model = build_model()
model.summary()


class PrintDoc(Callback):
    def on_epoch_end(self, epoch, logs=None):
        if epoch % 100 == 0:
            print('')
        print('.', end='')


EPOCHS = 1000

print(normed_train_data)
print(train_labels)

history = model.fit(
    normed_train_data, train_labels,
    epochs=EPOCHS, validation_split=0.2, verbose=0,
    callbacks=[PrintDoc()]
)

hist = pd.DataFrame(history.history)
hist['epoch'] = history.epoch
hist.tail()

Nagoya.csv

2021,1,1,2021/1/1,3,0,2.5,3.8,5.3,5.3,69,1019.2,1012.2,5.9,1.2,168
2021,1,2,2021/1/2,3.7,0,6.4,2.7,4.3,5.5,70,1020.8,1013.7,8.3,-1,92
2021,1,3,2021/1/3,4,0,1.2,2,10,6.1,75,1024.8,1017.7,7.9,1.5,152
2021,1,4,2021/1/4,5.6,0,6.7,2.6,5.3,6.4,72,1027.2,1020.2,11,2.7,99
2021,1,5,2021/1/5,5.3,0,1.1,1.8,9.3,6.5,74,1021.9,1014.9,8.6,2,161
2021,1,6,2021/1/6,6.5,0,5.4,2.9,8.5,5.8,61,1015.5,1008.6,10.7,3.9,103
2021,1,7,2021/1/7,3.3,0,4.5,4.9,9.8,4,52,1011.6,1004.6,8.1,-0.9,153
2021,1,8,2021/1/8,0,0,6.3,3.5,1.5,2.7,44,1013.2,1006.1,4.1,-2.2,92
2021,1,9,2021/1/9,0.5,0,8.7,2.4,1.5,2.7,43,1014.2,1007.1,5.8,-3.7,92
2021,1,10,2021/1/10,1.3,0,7.7,2.4,3.5,4.2,61,1020.1,1012.9,5.8,-1.8,92
2021,1,11,2021/1/11,2.2,0,2,1.8,8,5.2,72,1023.7,1016.6,6,-1,162
2021,1,12,2021/1/12,2.9,1.5,0.2,1.8,8.8,6.8,90,1016.4,1009.3,4.4,1.2,8
2021,1,13,2021/1/13,3.8,0,9.4,2.2,0.8,6,77,1018.1,1011.1,10.1,-0.6,92
2021,1,14,2021/1/14,6.3,0,7,2.6,2.3,6.6,71,1021.3,1014.3,12.3,0.6,99
2021,1,15,2021/1/15,7,0,9,1.5,2,6.5,67,1021.6,1014.6,14.2,1.1,98
2021,1,16,2021/1/16,6.8,0,3.6,2.2,6.8,7.9,80,1015.4,1008.5,12.5,2.7,234
2021,1,17,2021/1/17,4.7,0,1.1,3.2,9,6.2,73,1017.5,1010.5,9.1,2.4,184
2021,1,18,2021/1/18,4.2,0,9.3,3.4,3.5,5,63,1017,1010,8.6,0.1,92
2021,1,19,2021/1/19,3.1,0,6.8,5.5,4,4.5,59,1023.4,1016.3,6.5,-0.2,92
2021,1,20,2021/1/20,3.1,0,9.9,2.5,0.3,4,52,1029.7,1022.6,9.2,-2.1,98
2021,1,21,2021/1/21,5.4,0,8.9,1.5,3.5,4.7,55,1026.4,1019.3,13.1,-1.2,123
2021,1,22,2021/1/22,7.7,0.5,0,1.8,10,8.2,76,1019.6,1012.7,11.1,4.5,8
2021,1,23,2021/1/23,8.3,19,0,2.4,10,10.6,97,1017.1,1010.2,8.9,7.4,1
2021,1,24,2021/1/24,8.2,10.5,0,2.5,10,10.5,96,1018.4,1011.5,10.2,7,1
2021,1,25,2021/1/25,9,0,9.3,2.4,1.3,8.1,73,1024.7,1017.8,15.4,4.8,92
2021,1,26,2021/1/26,9.6,8,6.6,2,9,8.7,75,1021.6,1014.7,15.4,3.9,161
2021,1,27,2021/1/27,10.9,15,6.7,4.9,5.5,9.2,71,1013.7,1006.9,15.1,6.7,103
2021,1,28,2021/1/28,7.7,0,0,1.9,10,6.7,64,1010.4,1003.5,10.2,5.1,152
2021,1,29,2021/1/29,2.4,1.5,4.9,6.1,9.5,5.4,75,1011.7,1004.6,7.7,0,96
2021,1,30,2021/1/30,2.8,0.5,6.8,3.1,7,5.4,73,1022.6,1015.5,6.8,-1.2,109
2021,1,31,2021/1/31,5.7,0,9.7,3.2,3.8,5.7,65,1028,1021,10.9,2.2,92

補足情報（FW/ツールのバージョンなど）

python 3.9.0
tensflow 2.6.0

Answer 1

train_stats = train_dataset.describe()
train_stats = train_stats.transpose()

label = 'ave_temp'
train_labels = train_dataset.pop(label)
test_labels = test_dataset.pop(label)

上記を下記にしたところエラーは出なくなりました。

label = 'ave_temp'
train_labels = train_dataset.pop(label)
test_labels = test_dataset.pop(label)

train_stats = train_dataset.describe()
train_stats = train_stats.transpose()

Comments

[GraterBrown]

ありがとうございます。
参考にした記事をよく見たら
train_stats = train_dataset.describe()
train_stats.pop('MPG')
train_stats = train_stats.transpose()

train_labels = train_dataset.pop('MPG')
test_labels = test_dataset.pop('MPG')
となっていました。