TensorFlow requirements
TensorFlow requirements
- Development phase:- when you are coding and training a neural network.
- Runtime(or inference) phase:- when you are making predictions with a trained neural network.
- Development phase requirements:- windows, macOS, or Linuxcan use multiple Linux computer(locally or in the cloud ) for very large projects.
- Runtime phase supports:- computers running Windows, macOS or Linux.Linux servers running TensorFlow serving. google's cloud machine learning engine service. IOS or Android mobile apps.
GPU acceleration
Tensorflow can take advantage of NVIDIA-brand GPUs.
GPUs can greatly decrease neural network training times for large neural networks.
Using a GPU with TensorFlow requires installing additional software from NVIDIA(CUDA and cuDNN).
Programming language support
TensorFlow's core execution is written in c++ for speed. Python is the best supported and easiest language to use with TensorFlow.
Supervised learning
The branch of machine learning where the computer learn how to perform a function by looking at labelled training data.
Comments
Post a Comment