cuDNN, a GPU-accelerated library of primitives for deep neural networks.
The pre-configured and ready-to-use runtime environment for the Open Machine Learning Course, 2018. It includes Python 3.6, TensorFlow 1.4, Keras 2, XGBoost, LightGBM and Vowpal Wabbit. The stack also includes CUDA and cuDNN, and is optimized for running on NVidia GPU.
mlcourse_open-course:2018, cuda:8.0.61, cudnn:7.0.5, cuda_only-nvidia_drivers:384.111