README.Rmd

---
title: "deeplyr"
output: github_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = T, eval = T, message = F, warning = F, error = F)
```

```{r, echo = F, results='asis'}
library(badger)
cat(
	badge_code_size("systats/deeplyr"),
	badge_last_commit("systats/deeplyr")
)
```

`deeplyr` provides a standard model interface for some machine learning libraries. Internally the package is managed by R6 classes and hardhat allowing for cross-validation or genetic hyper parameter tuning. Every model run can be saved including meta, params, metrics, process, model, and preds either as .rds or .json.

Have already full support (fit, predict, save, load)

* `keras`
* `h2o`
* `xgboost`