Keras堅持第一紀元

Question

我試圖測試Keras和TensorFlow是否在最新的Mojave上運行我的MacBook Pro，內存為32GB，顯然它不是！

我嘗試在一個單獨的新環境中安裝它們，它工作正常，但我不明白為什么它不能在我的基礎（根）環境中工作。

from keras.models import Sequential
from keras.layers import Dense
from keras.utils import to_categorical

from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.datasets import make_regression

X, y = make_regression(n_samples=1000, n_features=20)
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)

ss = StandardScaler()
X_train_sc = ss.fit_transform(X_train)
X_test_sc = ss.transform(X_test)

model = Sequential()
model.add(Dense(32, input_dim=20, activation='relu'))
model.add(Dense(8, activation='relu'))
model.add(Dense(1))

model.compile(loss='mean_squared_error', optimizer='adam')

model.fit(X_train_sc, y_train, validation_data=(X_test_sc, y_test), epochs=10, batch_size=32)

我希望得到這個結果，我在干凈的環境中做了這個結果：

WARNING:tensorflow:From /Users/Hovanes/anaconda3/envs/clean/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.cast instead.
Train on 750 samples, validate on 250 samples
Epoch 1/10
2019-04-26 19:04:32.220021: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
750/750 [==============================] - 0s 235us/step - loss: 30460.2991 - val_loss: 30451.6543
Epoch 2/10
750/750 [==============================] - 0s 23us/step - loss: 30384.4905 - val_loss: 30375.6350
Epoch 3/10
750/750 [==============================] - 0s 22us/step - loss: 30292.1559 - val_loss: 30280.5673
Epoch 4/10
750/750 [==============================] - 0s 23us/step - loss: 30162.1524 - val_loss: 30141.1293
Epoch 5/10
750/750 [==============================] - 0s 22us/step - loss: 29971.8937 - val_loss: 29918.3467
Epoch 6/10
750/750 [==============================] - 0s 23us/step - loss: 29689.4520 - val_loss: 29591.1545
Epoch 7/10
750/750 [==============================] - 0s 22us/step - loss: 29266.3404 - val_loss: 29122.6358
Epoch 8/10
750/750 [==============================] - 0s 22us/step - loss: 28671.3374 - val_loss: 28470.9937
Epoch 9/10
750/750 [==============================] - 0s 23us/step - loss: 27898.4042 - val_loss: 27585.4375
Epoch 10/10
750/750 [==============================] - 0s 22us/step - loss: 26869.9945 - val_loss: 26530.5343
<keras.callbacks.History object at 0x136d07630>

但相反，我只有這個：

WARNING:tensorflow:From /Users/Hovanes/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.cast instead.
Train on 750 samples, validate on 250 samples
Epoch 1/10

我使用完全相同的安裝方法（pip）在完全相同的計算機上運行完全相同的代碼。

任何和所有的幫助將不勝感激！

Answer 1

無法復制錯誤，我在manjaro中使用了python 3.6.7和3.7.3。 要下載我使用的軟件包：

conda install -c conda-forge jupyterlab scikit-learn keras

你能告訴我你是如何安裝這些庫的嗎？ 否則，您可以嘗試使用上面的命令來安裝庫。

如果它不起作用可能是別的。

Answer 2

所以，問題是pip install ...我卸載了所有東西並通過conda-forge安裝它，它終於工作了。 希望它能幫助那些將來遇到這個問題的人！