MXNet™
Prijs van MXNet™: gratis
Versie: 0.11.0
MXNet™ is an open source library designed for deep learning, with which can be used to train and deploy deep neural networks on any devices. The abbreviation of the name of the library stands for two words: mix and maximize. MXNet™ library allows taking advantage of combined power of declarative and imperative programming, which is MXNet™ distinctive advantage.
MXNet™ library includes a dynamic dependency scheduler that automatically parallels character and imperative operations.
It is worth mentioning some other advantages of the library such as portability and ease of use. It can be easily scaled to multiple GPUs, servers and clusters.
MXNet™ library perfectly fits for distributed computing. MXNet™ supports a whole range of languages such as R, Python, Julia, etc. We encourage you to take an advantage of using MXNet™ on rented GPU servers in LeaderGPU® service.
1. Install CUDA® 9.0
2. Install cuDNN 7 for CUDA® 9.0
3. Install MXNet™ with GPU support using CUDA® 9.2
$ pip install mxnet-cu92
4. Install Graphviz.
sudo apt-get install graphviz
pip install graphviz
5. Validate Mxnet™
python
>>> import mxnet as mx
>>> a = mx.nd.ones((2, 3), mx.gpu())
>>> b = a * 2 + 1
>>> b.asnumpy()
array([[ 3., 3., 3.],
[3., 3., 3.]], dtype=float32)
Categorie
Machine learning
Branche
Automotive, Aerospace, Electronic, Biotechnology, Finance
Licentie
Opensource/Free