ggml_matrix-class

%*%,ggml_matrix,ANY-method

%*%,ANY,ggml_matrix-method

%*%,ggml_matrix,ggml_matrix-method

crossprod,ggml_matrix-method

tcrossprod,ggml_matrix-method

dim,ggml_matrix-method

show,ggml_matrix-method

An S4 class holding an ordinary numeric matrix, tagged so that <code>%*%</code>,
<code><a href="/link/crossprod()?package=ggmlR&version=0.8.1" data-mini-rdoc="ggmlR::crossprod()">crossprod()</a></code> and <code><a href="/link/tcrossprod()?package=ggmlR&version=0.8.1" data-mini-rdoc="ggmlR::tcrossprod()">tcrossprod()</a></code> run on the GPU (see
<code>ggml_matmul</code>). Wrap a matrix with <code>as_gpu_matrix</code>.

Provides 'R' bindings to the 'GGML' tensor library for machine
learning, optimized for 'Vulkan' GPU acceleration with a transparent CPU
fallback. The package features a 'Keras'-like sequential API and a
'PyTorch'-style 'autograd' engine for building, training, and deploying
neural networks. Key capabilities include high-performance 5D tensor
operations, 'f16' precision, and efficient quantization. It supports
native 'ONNX' model import (50+ operators) and 'GGUF' weight loading
from the 'llama.cpp' and 'Hugging Face' ecosystems. Designed for
zero-overhead inference via dedicated weight buffering, it integrates
seamlessly as a 'parsnip' engine for 'tidymodels' and provides
first-class learners for the 'mlr3' framework.
See <https://github.com/ggml-org/ggml> for more information about the
underlying library.

Yuri Baramykov

ggmlR

'GGML' Tensor Operations for Machine Learning

ggml_matrix-class function

<dl><dt>x, y</dt>
<dd>Operands; at least one is a <code>ggml_matrix</code> (the other is
coerced with <code>as.matrix</code>). For <code>crossprod</code>/<code>tcrossprod</code>,
<code>y</code> may be missing.</dd>
<dt>object</dt>
<dd>A <code>ggml_matrix</code> (for the <code>show</code> method).</dd></dl>

Arguments

<dl>
<dt><code>data</code></dt>
<dd>The underlying numeric matrix.</dd></dl>
<dt><code>device</code></dt>
<dd>Dispatch preference: <code>"auto"</code>, <code>"gpu"</code> or <code>"cpu"</code>.</dd>
<dt><code>prec</code></dt>
<dd>Precision of the GPU path: <code>"f32"</code> or <code>"f16"</code>.</dd>

Slots

An S4 class holding an ordinary numeric matrix, tagged so that <code>%*%</code>,
<code><a href='https://rdrr.io/r/base/crossprod.html'>crossprod()</a></code> and <code><a href='https://rdrr.io/r/base/crossprod.html'>tcrossprod()</a></code> run on the GPU (see
<code>ggml_matmul</code>). Wrap a matrix with <code>as_gpu_matrix</code>.

A GPU-backed matrix wrapper — ggml_matrix-class

<dl>

<dt>x, y</dt>
<dd>Operands; at least one is a <code>ggml_matrix</code> (the other is
coerced with <code>as.matrix</code>). For <code>crossprod</code>/<code>tcrossprod</code>,
<code>y</code> may be missing.</dd>


<dt>object</dt>
<dd>A <code>ggml_matrix</code> (for the <code>show</code> method).</dd>

</dl>

<dl>
<dt><code>data</code></dt>
<dd>The underlying numeric matrix.</dd>


<dt><code>device</code></dt>
<dd>Dispatch preference: <code>"auto"</code>, <code>"gpu"</code> or <code>"cpu"</code>.</dd>


<dt><code>prec</code></dt>
<dd>Precision of the GPU path: <code>"f32"</code> or <code>"f16"</code>.</dd>


</dl>

ggml_matrix-class: A GPU-backed matrix wrapper

Description

Usage

Value

Arguments

Slots