The ultimate launch of TensorFlow v1.3 is now accessible. This launch of TensorFlow marks the preliminary availability of a number of canned estimators, together with:
- DNNClassifier
- DNNRegressor
- LinearClassifier
- LinearRegressor
- DNNLinearCombinedClassifier
- DNNLinearCombinedRegressor.
The tfestimators bundle gives a excessive degree R interface for these estimators.
Full particulars on the discharge of TensorFlow v1.3 can be found right here: https://github.com/tensorflow/tensorflow/releases/tag/v1.3.0
You’ll be able to replace your R set up of TensorFlow utilizing the install_tensorflow
perform:
library(tensorflow)
install_tensorflow()
Be aware that you just must also present any choices utilized in your authentic set up (e.g. methodology = "conda"
, model = "gpu"
, and so on. )
cuDNN 6.0
TensorFlow v1.3 is constructed in opposition to model 6.0 of the cuDNN library from NVIDIA. Earlier variations have been constructed in opposition to cuDNN v5.1, so for installations working the GPU model of TensorFlow this implies that you’ll want to put in an up to date model of cuDNN together with TensorFlow v1.3.
Up to date set up directions can be found right here: https://tensorflow.rstudio.com/tensorflow/installation_gpu.html.
Model 1.4 of TensorFlow is predicted emigrate once more to model 7.0 of cuDNN.
Reuse
Textual content and figures are licensed underneath Inventive Commons Attribution CC BY 4.0. The figures which have been reused from different sources do not fall underneath this license and might be acknowledged by a word of their caption: “Determine from …”.
Quotation
For attribution, please cite this work as
Allaire (2017, Aug. 17). Posit AI Weblog: TensorFlow v1.3 Launched. Retrieved from https://blogs.rstudio.com/tensorflow/posts/2017-08-17-tensorflow-v13-released/
BibTeX quotation
@misc{allaire2017tensorflow, writer = {Allaire, J.J.}, title = {Posit AI Weblog: TensorFlow v1.3 Launched}, url = {https://blogs.rstudio.com/tensorflow/posts/2017-08-17-tensorflow-v13-released/}, 12 months = {2017} }