Info
| Version: | 0.4.11 |
| Author(s): | Deep.Net Developers |
| Last Update: | Tuesday, May 8, 2018 |
| .NET Fiddle: | Create the first Fiddle |
| Project Url: | http://www.deepml.net/Tensor |
| NuGet Url: | https://www.nuget.org/packages/Tensor |
Install
Install-Package Tensor
dotnet add package Tensor
paket add Tensor
Tensor Download (Unzip the "nupkg" after downloading)
Dependencies
- FSharp.Core(>= 4.3.4)
- HDF.PInvoke.NETStandard(>= 1.10.200)
- ManagedCuda-CUBLAS.NETStandard(>= 9.1.300)
- ManagedCuda-NVRTC.NETStandard(>= 9.1.300)
- ManagedCuda.NETStandard(>= 9.1.300)
- System.Numerics.Vectors(>= 4.4.0)
- System.Reflection.Emit(>= 4.3.0)
Tags
Core features:
- n-dimensional arrays (tensors) in host memory or on CUDA GPUs
- element-wise operations (addition, multiplication, absolute value, etc.)
- basic linear algebra operations (dot product, SVD decomposition, matrix inverse, etc.)
- reduction operations (sum, product, average, maximum, arg max, etc.)
- logic operations (comparision, and, or, etc.)
- views, slicing, reshaping, broadcasting (similar to NumPy)
- scatter and gather by indices
- standard functional operations (map, fold, etc.)
Data exchange:
- read/write support for HDF5 (.h5)
- interop with standard F# types (Seq, List, Array, Array2D, Array3D, etc.)
Performance:
- host: SIMD and BLAS accelerated operations
- by default Intel MKL is used (shipped with NuGet package)
- other BLASes (OpenBLAS, vendor-specific) can be selected by configuration option
- CUDA GPU: all operations performed locally on GPU and cuBLAS used for matrix operations
Requirements:
- Linux, MacOS or Windows on x64
- Linux requires libgomp.so.1 installed.
Additional algorithms are provided in the Tensor.Algorithm package.