teldufalsari/kmath
2
0
Fork 0
forked from kscience/kmath
Code Pull Requests 1 Activity

Corrected readme file

Browse Source
This commit is contained in:
Roland Grinis 2021-01-16 20:29:47 +00:00
parent ed4ac2623d
commit e5205d5afd
1 changed files with 22 additions and 11 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

21
kmath-torch/README.md
View File

@ -1,9 +1,12 @@
# LibTorch extension (`kmath-torch`)
This is a `Kotlin/Native` module, with only `linuxX64` supported so far. This library wraps some of the [PyTorch C++ API](https://pytorch.org/cppdocs), focusing on integrating `Aten` & `Autograd` with `KMath`.
This is a `Kotlin/Native` module, with only `linuxX64` supported so far. This library wraps some of
the [PyTorch C++ API](https://pytorch.org/cppdocs), focusing on integrating `Aten` & `Autograd` with `KMath`.
## Installation
To install the library, you have to build & publish locally `kmath-core`, `kmath-memory` with `kmath-torch`:
```
./gradlew -q :kmath-core:publishToMavenLocal :kmath-memory:publishToMavenLocal :kmath-torch:publishToMavenLocal
```
@ -13,6 +16,7 @@ This builds `ctorch`, a C wrapper for `LibTorch` placed inside:
`~/.konan/third-party/kmath-torch-0.2.0-dev-4/cpp-build`
You will have to link against it in your own project. Here is an example of build script for a standalone application:
```kotlin
//build.gradle.kts
plugins {
@ -59,6 +63,7 @@ kotlin {
}
}
```
```kotlin
//settings.gradle.kts
pluginManagement {
@ -76,7 +81,9 @@ pluginManagement {
## Usage
Tensors are implemented over the `MutableNDStructure`. They can only be instantiated through provided factory methods and require scoping:
Tensors are implemented over the `MutableNDStructure`. They can only be instantiated through provided factory methods
and require scoping:
```kotlin
TorchTensorRealAlgebra {
@ -94,23 +101,27 @@ TorchTensorRealAlgebra {
println(gpuRealTensor)
}
```
Enjoy a high performance automatic differentiation engine:
```kotlin
TorchTensorRealAlgebra {
val dim = 10
val device = TorchDevice.TorchCPU //or TorchDevice.TorchCUDA(0)
val x = randNormal(shape = intArrayOf(dim), device = device)
// x is the variable
x.requiresGrad = true
val X = randNormal(shape = intArrayOf(dim, dim), device = device)
val Q = X + X.transpose(0, 1)
val mu = randNormal(shape = intArrayOf(dim), device = device)
// expression to differentiate w.r.t. x
val f = 0.5 * (x dot (Q dot x)) + (mu dot x) + 25.3
val f = x.withGrad {
0.5 * (x dot (Q dot x)) + (mu dot x) + 25.3
}
// value of the gradient at x
val gradf = f grad x
// value of the hessian at x
val hessf = f hess x
}
```