In AWS SageMaker there are several Conda environments available based on different ML frameworks:
conda info -e
I would like to know the difference between a tensorflow
and an amazonei_tensorflow
environment.
The amazonei
environments are for use with the Amazon Elastic Inference accelerators. You can see this mentioned in the DLAMI README file:
for TensorFlow(+Keras2) with Python3 (CUDA 9.0 and Intel MKL-DNN) ______________
____ source activate tensorflow_p36
for Tensorflow(+Amazon Elastic Inference) with Python3 _________________________
____ source activate amazonei_tensorflow_p36