I work with Tensor Flow on python 2.7 and I get the following warnings. What are they connected with, I do not understand.

2017-09-06 22: 28: 21.659388: W tensorflow / core / platform / cpu_feature_guard.cc: 45] SSE4.1 CPU computations.

2017-09-06 22: 28: 21.659414: W tensorflow / core / platform / cpu_feature_guard.cc: 45] SSE4.2 cpu computations.

2017-09-06 22: 28: 21.659419: W tensorflow / core / platform / cpu_feature_guard.cc: 45] AVX instructions computations.

2017-09-06 22: 28: 21.659423: W tensorflow / core / platform / cpu_feature_guard.cc: 45] AVX2 instructions for computations.

2017-09-06 22: 28: 21.659427: W tensorflow / core / platform / cpu_feature_guard.cc: 45] FMA instructions and computations.

Model example:

def baseline_model(): model = Sequential() model.add(Dense(8, input_dim=4, activation='relu')) model.add(Dense(3, activation='softmax')) model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy']) return model estimator = KerasClassifier(build_fn=baseline_model, epochs=200, batch_size=5, verbose=0) kfold = KFold(n_splits=10, shuffle=True, random_state=seed) results = cross_val_score(estimator, X, dummy_y, cv=kfold)

    1 answer 1

    They say that TensorFlow, compiled from sources, can work faster on your machine. These instructions are not included by default in the available assemblies. I think that they are compatible with a large number of processors.

    To disable these warnings, follow these steps:

     import os os.environ['TF_CPP_MIN_LOG_LEVEL']='2' import tensorflow as tf
    • Or. as an option, you can rebuild tensorflow from source. However, if all the training goes to the GPU, there will be practically no performance gain. - velikodniy