The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.

Mars Runtime Basics

Overview

marsruntime provides a portable R surface for validated Mars ModelSpec artifacts. It can validate, replay, and inspect portable artifacts through the same package interface that the release docs and package manual describe.

Loading a model specification

library(marsruntime)

spec <- list(
  spec_version = "1.0",
  basis_terms = list(list(kind = "constant")),
  coefficients = list(1),
  feature_schema = list(n_features = 0)
)

validate_model_spec(spec)

Replaying predictions

rows <- matrix(numeric(0), nrow = 1, ncol = 0)
design_matrix(spec, rows)
#>      [,1]
#> [1,]    1
predict_model(spec, rows)
#> [1] 1

Training

When the Rust runtime helper is available, fit_model() can fit a portable specification from feature and response data. The actual training path is runtime-dependent, so this vignette shows the call without executing it.

fit_model(
  x = matrix(c(0, 1, 2), ncol = 1),
  y = c(1, 3, 5),
  max_terms = 5,
  max_degree = 1,
  penalty = 3.0
)

Validation

For source-tree validation, run:

Rscript tests/conformance.R
Rscript tests/training.R

For package publication validation, run:

R CMD build .
R CMD check --no-manual --as-cran marsruntime_*.tar.gz
R CMD Rd2pdf .

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.