**StreamExecutor device (0): Host, Default Version** in Tensorflow

user_14520701 0
Viewed 1546

I am trying to execute a tensorflow script on my local computer. But the following warning appears. It would be really helpful if someone can specify what the common cause for this warning is.

2020-10-26 14:04:34.690753: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN)to use the following CPU instructions in performance-critical operations:  AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2020-10-26 14:04:34.707717: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7fadab4339f0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-10-26 14:04:34.707751: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version

I tried tracing my code and this is the line that gives the error :

model = load_model(MODEL_PATH)

and I import load_model using

from keras.models import load_model
tensorflow deep-learning machine-learning python
1 Answers

The main reason for the above warning is the Tensorflow library is not optimized for your particular architecture which Tensorflow was originally compiled on different architecture machine.
You can ignore this warning but you won't be getting maximum performance from the library.

To fix this warning and get maximum performance, you need to build Tensorflow on your machine.

You can follow this documentation to build Tensorflow from source for Linux and macOS for Windows you can follow this documentation.

user_142906970
Related