diff --git a/.Rbuildignore b/.Rbuildignore
index 999be025..aa881e6b 100644
--- a/.Rbuildignore
+++ b/.Rbuildignore
@@ -8,3 +8,5 @@
 ^pkgdown$
 ^doc$
 ^Meta$
+^cran-comments\.md$
+^CRAN-RELEASE$
diff --git a/CRAN-RELEASE b/CRAN-RELEASE
new file mode 100644
index 00000000..04cc61c9
--- /dev/null
+++ b/CRAN-RELEASE
@@ -0,0 +1,2 @@
+This package was submitted to CRAN on 2021-06-16.
+Once it is accepted, delete this file and tag the release (commit 1d00ae2).
diff --git a/DESCRIPTION b/DESCRIPTION
index 2dc1f294..9206f6d2 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: luz
 Title: Higher Level 'API' for 'torch'
-Version: 0.0.0.9000
+Version: 0.1.0
 Authors@R: c(
     person("Daniel", "Falbel", email = "daniel@rstudio.com", role = c("aut", "cre", "cph")),
     person(family = "RStudio", role = c("cph"))
@@ -12,9 +12,8 @@ Description: A high level interface for 'torch' providing utilities to reduce th
     Howard et al. (2020) <arXiv:2002.04688>, 'Keras' by Chollet et al. (2015) and 
     'Pytorch Lightning' by Falcon et al. (2019) <doi:10.5281/zenodo.3828935>.
 License: MIT + file LICENSE
-URL: https://mlverse.github.io/luz, https://github.com/mlverse/luz
+URL: https://mlverse.github.io/luz/, https://github.com/mlverse/luz
 Encoding: UTF-8
-LazyData: true
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.1.1
 Imports: 
diff --git a/R/callbacks-interrupt.R b/R/callbacks-interrupt.R
index 6270fa58..fe9d3e1b 100644
--- a/R/callbacks-interrupt.R
+++ b/R/callbacks-interrupt.R
@@ -10,6 +10,12 @@ NULL
 #' @note In general you don't need to use these callback by yourself because it's always
 #'   included by default in [fit.luz_module_generator()].
 #'
+#' @examples
+#' interrupt_callback <- luz_callback_interrupt()
+#'
+#' @returns
+#' A `luz_callback`
+#'
 #' @family luz_callbacks
 #' @export
 luz_callback_interrupt <- luz_callback(
diff --git a/R/callbacks-profile.R b/R/callbacks-profile.R
index 95782834..f8cd44e8 100644
--- a/R/callbacks-profile.R
+++ b/R/callbacks-profile.R
@@ -16,6 +16,12 @@
 #' @note In general you don't need to use these callback by yourself because it's always
 #'   included by default in [fit.luz_module_generator()].
 #'
+#' @examples
+#' profile_callback <- luz_callback_profile()
+#'
+#' @returns
+#' A `luz_callback`
+#'
 #' @family luz_callbacks
 #' @export
 luz_callback_profile <- luz_callback(
diff --git a/R/callbacks.R b/R/callbacks.R
index a2cf5ef5..ca5f70a9 100644
--- a/R/callbacks.R
+++ b/R/callbacks.R
@@ -86,6 +86,10 @@ luz_callback <- function(name = NULL, ..., private = NULL, active = NULL, parent
 #' @note Printing can be disabled by passing `verbose=FALSE` to [fit.luz_module_generator()].
 #'
 #' @family luz_callbacks
+#'
+#' @returns
+#' A `luz_callback`
+#'
 #' @export
 luz_callback_progress <- luz_callback(
   "progress_callback",
@@ -203,6 +207,10 @@ luz_callback_progress <- luz_callback(
 #' used by default in [fit.luz_module_generator()].
 #'
 #' @family luz_callbacks
+#'
+#' @returns
+#' A `luz_callback`
+#'
 #' @export
 luz_callback_metrics <- luz_callback(
   "metrics_callback",
@@ -272,6 +280,9 @@ luz_callback_metrics <- luz_callback(
 #' @note In general you won't need to explicitly use the metrics callback as it's
 #' used by default in [fit.luz_module_generator()].
 #'
+#' @returns
+#' A `luz_callback`
+#'
 #' @family luz_callbacks
 #' @export
 luz_callback_train_valid <- luz_callback(
diff --git a/R/metrics.R b/R/metrics.R
index e17156e0..1216edb8 100644
--- a/R/metrics.R
+++ b/R/metrics.R
@@ -96,6 +96,11 @@ luz_metric <- function(name = NULL, ..., private = NULL, active = NULL,
 #' metric$compute()
 #' }
 #' @export
+#'
+#'
+#' @returns
+#' Returns new Luz metric.
+#'
 #' @family luz_metrics
 luz_metric_accuracy <- luz_metric(
   abbrev = "Acc",
@@ -131,6 +136,10 @@ luz_metric_accuracy <- luz_metric(
 #' metric$update(torch_rand(100), torch::torch_randint(0, 1, size = 100))
 #' metric$compute()
 #' }
+#'
+#' @returns
+#' Returns new Luz metric.
+#'
 #' @family luz_metrics
 #' @export
 luz_metric_binary_accuracy <- luz_metric(
@@ -172,6 +181,8 @@ luz_metric_binary_accuracy <- luz_metric(
 #' metric$update(torch_randn(100), torch::torch_randint(0, 1, size = 100))
 #' metric$compute()
 #' }
+#' @returns
+#' Returns new Luz metric.
 #'
 #' @family luz_metrics
 #' @export
@@ -240,6 +251,8 @@ luz_metric_loss_average <- luz_metric(
 #' metric$update(torch_randn(100), torch_randn(100))
 #' metric$compute()
 #' }
+#' @returns
+#' Returns new Luz metric.
 #'
 #' @family luz_metrics
 #' @export
@@ -289,6 +302,10 @@ luz_metric_mse <- luz_metric(
 #' Computes the root mean squared error.
 #'
 #' @family luz_metrics
+#'
+#' @returns
+#' Returns new Luz metric.
+#'
 #' @export
 luz_metric_rmse <- luz_metric(
   inherit = luz_metric_mse,
diff --git a/R/module-plot.R b/R/module-plot.R
index 83293a14..00635944 100644
--- a/R/module-plot.R
+++ b/R/module-plot.R
@@ -5,3 +5,5 @@ plot.luz_module_fitted <- function(x, ...) {
   p <- p + ggplot2::geom_point() + ggplot2::geom_line()
   p + ggplot2::facet_grid(metric ~ set, scales = "free_y")
 }
+
+globalVariables(c("epoch", "value"))
diff --git a/R/module-print.R b/R/module-print.R
index d4151fe1..2d9f73d9 100644
--- a/R/module-print.R
+++ b/R/module-print.R
@@ -43,3 +43,5 @@ print.luz_module_fitted <- function(x, ...) {
   print(x$model)
 }
 
+
+
diff --git a/R/module.R b/R/module.R
index fa9379b9..dc56aba5 100644
--- a/R/module.R
+++ b/R/module.R
@@ -16,6 +16,9 @@
 #' @param metrics (`list`, optional) A list of metrics to be tracked during
 #' the training procedure.
 #'
+#' @returns
+#' A luz module that can be trained with [fit()].
+#'
 #' @family training
 #'
 #' @export
@@ -68,6 +71,9 @@ setup <- function(module, loss = NULL, optimizer = NULL, metrics = NULL) {
 #'
 #' @family set_hparam
 #'
+#' @returns
+#' The same luz module
+#'
 #' @export
 set_hparams <- function(module, ...) {
   hparams <- rlang::list2(...)
@@ -87,6 +93,9 @@ set_hparams <- function(module, ...) {
 #' `optim_adam` function is called with `optim_adam(parameters, lr=0.1)` when fitting
 #' the model.
 #'
+#' @returns
+#' The same luz module
+#'
 #' @family set_hparam
 #' @export
 set_opt_hparams <- function(module, ...) {
@@ -132,9 +141,12 @@ get_opt_hparams <- function(module) {
 #'
 #' @param ... Currently unused,
 #'
+#' @returns
+#' A fitted object that can be saved with [luz_save()] and can be printed with
+#' [print()] and plotted with [plot()].
+#'
 #' @importFrom generics fit
 #' @export
-#'
 fit.luz_module_generator <- function(object, data, epochs = 10, callbacks = NULL,
                                      valid_data = NULL, accelerator = NULL,
                                      verbose = NULL, ...) {
diff --git a/README.md b/README.md
index 7b13bf6e..9c80eb5d 100644
--- a/README.md
+++ b/README.md
@@ -3,7 +3,7 @@
 <!-- badges: start -->
 [![R-CMD-check](https://github.com/mlverse/luz/workflows/R-CMD-check/badge.svg)](https://github.com/mlverse/luz/actions)
 [![Codecov test coverage](https://codecov.io/gh/mlverse/luz/branch/master/graph/badge.svg)](https://codecov.io/gh/mlverse/luz?branch=master)
-[![Discord](https://img.shields.io/discord/837019024499277855?logo=discord)](https://discord.gg/s3D5cKhBkx)
+[![Discord](https://img.shields.io/discord/837019024499277855?logo=discord)](https://discord.com/invite/s3D5cKhBkx)
 <!-- badges: end -->
 
 luz is a higher level API for torch providing abstractions to allow for much less verbose training loops.
diff --git a/cran-comments.md b/cran-comments.md
new file mode 100644
index 00000000..0b524c9f
--- /dev/null
+++ b/cran-comments.md
@@ -0,0 +1 @@
+First release.
diff --git a/man/fit.luz_module_generator.Rd b/man/fit.luz_module_generator.Rd
index 6b7056a8..0ac48981 100644
--- a/man/fit.luz_module_generator.Rd
+++ b/man/fit.luz_module_generator.Rd
@@ -43,6 +43,10 @@ to the console.}
 
 \item{...}{Currently unused,}
 }
+\value{
+A fitted object that can be saved with \code{\link[=luz_save]{luz_save()}} and can be printed with
+\code{\link[=print]{print()}} and plotted with \code{\link[=plot]{plot()}}.
+}
 \description{
 Fit a \code{nn_module}
 }
diff --git a/man/luz_callback_interrupt.Rd b/man/luz_callback_interrupt.Rd
index 77c9553d..5d7c5369 100644
--- a/man/luz_callback_interrupt.Rd
+++ b/man/luz_callback_interrupt.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_callback_interrupt()
 }
+\value{
+A \code{luz_callback}
+}
 \description{
 Adds a handler that allows interrupting the training loop using \code{ctrl + C}.
 Also registers a \code{on_interrupt} breakpoint so users can register callbacks to
@@ -14,6 +17,10 @@ be run on training loop interruption.
 \note{
 In general you don't need to use these callback by yourself because it's always
 included by default in \code{\link[=fit.luz_module_generator]{fit.luz_module_generator()}}.
+}
+\examples{
+interrupt_callback <- luz_callback_interrupt()
+
 }
 \seealso{
 Other luz_callbacks: 
diff --git a/man/luz_callback_metrics.Rd b/man/luz_callback_metrics.Rd
index 5f03ff8f..2a70cb77 100644
--- a/man/luz_callback_metrics.Rd
+++ b/man/luz_callback_metrics.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_callback_metrics()
 }
+\value{
+A \code{luz_callback}
+}
 \description{
 Tracks metrics passed to \code{\link[=setup]{setup()}} during training and validation.
 }
diff --git a/man/luz_callback_profile.Rd b/man/luz_callback_profile.Rd
index a3689251..04cf8119 100644
--- a/man/luz_callback_profile.Rd
+++ b/man/luz_callback_profile.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_callback_profile()
 }
+\value{
+A \code{luz_callback}
+}
 \description{
 Computes the times for high-level operations in the training loops.
 }
@@ -24,6 +27,10 @@ the model step. (not including data acquisition and preprocessing)
 \note{
 In general you don't need to use these callback by yourself because it's always
 included by default in \code{\link[=fit.luz_module_generator]{fit.luz_module_generator()}}.
+}
+\examples{
+profile_callback <- luz_callback_profile()
+
 }
 \seealso{
 Other luz_callbacks: 
diff --git a/man/luz_callback_progress.Rd b/man/luz_callback_progress.Rd
index 7e266e74..b0845df7 100644
--- a/man/luz_callback_progress.Rd
+++ b/man/luz_callback_progress.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_callback_progress()
 }
+\value{
+A \code{luz_callback}
+}
 \description{
 Responsible for printing progress during training.
 }
diff --git a/man/luz_callback_train_valid.Rd b/man/luz_callback_train_valid.Rd
index 10b9fc70..f2b09439 100644
--- a/man/luz_callback_train_valid.Rd
+++ b/man/luz_callback_train_valid.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_callback_train_valid()
 }
+\value{
+A \code{luz_callback}
+}
 \description{
 Switches important flags for training and evaluation modes.
 }
diff --git a/man/luz_metric_accuracy.Rd b/man/luz_metric_accuracy.Rd
index 60cf76f5..6c828736 100644
--- a/man/luz_metric_accuracy.Rd
+++ b/man/luz_metric_accuracy.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_metric_accuracy()
 }
+\value{
+Returns new Luz metric.
+}
 \description{
 Computes accuracy for multi-class classification problems.
 }
diff --git a/man/luz_metric_binary_accuracy.Rd b/man/luz_metric_binary_accuracy.Rd
index ce37c50e..c3b3adfc 100644
--- a/man/luz_metric_binary_accuracy.Rd
+++ b/man/luz_metric_binary_accuracy.Rd
@@ -9,6 +9,9 @@ luz_metric_binary_accuracy(threshold = 0.5)
 \arguments{
 \item{threshold}{value used to classifiy observations between 0 and 1.}
 }
+\value{
+Returns new Luz metric.
+}
 \description{
 Computes the accuracy for binary classification problems where the
 model returns probabilities. Commonly used when the loss is \code{\link[torch:nn_bce_loss]{torch::nn_bce_loss()}}.
@@ -21,6 +24,7 @@ metric <- metric$new()
 metric$update(torch_rand(100), torch::torch_randint(0, 1, size = 100))
 metric$compute()
 }
+
 }
 \seealso{
 Other luz_metrics: 
diff --git a/man/luz_metric_binary_accuracy_with_logits.Rd b/man/luz_metric_binary_accuracy_with_logits.Rd
index 83c15b77..240b7d35 100644
--- a/man/luz_metric_binary_accuracy_with_logits.Rd
+++ b/man/luz_metric_binary_accuracy_with_logits.Rd
@@ -9,6 +9,9 @@ luz_metric_binary_accuracy_with_logits(threshold = 0.5)
 \arguments{
 \item{threshold}{value used to classifiy observations between 0 and 1.}
 }
+\value{
+Returns new Luz metric.
+}
 \description{
 Computes accuracy for binary classification problems where the model
 return logits. Commonly used together with \code{\link[torch:nn_bce_with_logits_loss]{torch::nn_bce_with_logits_loss()}}.
@@ -25,7 +28,6 @@ metric <- metric$new()
 metric$update(torch_randn(100), torch::torch_randint(0, 1, size = 100))
 metric$compute()
 }
-
 }
 \seealso{
 Other luz_metrics: 
diff --git a/man/luz_metric_mae.Rd b/man/luz_metric_mae.Rd
index ed34603b..36b74fb2 100644
--- a/man/luz_metric_mae.Rd
+++ b/man/luz_metric_mae.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_metric_mae()
 }
+\value{
+Returns new Luz metric.
+}
 \description{
 Computes the mean absolute error.
 }
@@ -17,7 +20,6 @@ metric <- metric$new()
 metric$update(torch_randn(100), torch_randn(100))
 metric$compute()
 }
-
 }
 \seealso{
 Other luz_metrics: 
diff --git a/man/luz_metric_rmse.Rd b/man/luz_metric_rmse.Rd
index 97d3d1cd..ea28544a 100644
--- a/man/luz_metric_rmse.Rd
+++ b/man/luz_metric_rmse.Rd
@@ -6,6 +6,9 @@
 \usage{
 luz_metric_rmse()
 }
+\value{
+Returns new Luz metric.
+}
 \description{
 Computes the root mean squared error.
 }
diff --git a/man/set_hparams.Rd b/man/set_hparams.Rd
index fe3ccd8a..028dc6f0 100644
--- a/man/set_hparams.Rd
+++ b/man/set_hparams.Rd
@@ -12,6 +12,9 @@ set_hparams(module, ...)
 \item{...}{The parameters set here will be used to initialize the \code{nn_module}, ie they
 are passed unchanged to the \code{initialize} method of the base \code{nn_module}.}
 }
+\value{
+The same luz module
+}
 \description{
 This function is used to define hyper-parameters before calling \code{fit} for
 \code{luz_modules}.
diff --git a/man/set_opt_hparams.Rd b/man/set_opt_hparams.Rd
index 2e8c36c1..9ac0fc83 100644
--- a/man/set_opt_hparams.Rd
+++ b/man/set_opt_hparams.Rd
@@ -14,6 +14,9 @@ For example, if your optimizer is \code{optim_adam} and you pass \code{lr=0.1},
 \code{optim_adam} function is called with \code{optim_adam(parameters, lr=0.1)} when fitting
 the model.}
 }
+\value{
+The same luz module
+}
 \description{
 This function is used to define hyper-parameters for the optimizer initialization
 method.
diff --git a/man/setup.Rd b/man/setup.Rd
index c2b9d01d..22d6da9f 100644
--- a/man/setup.Rd
+++ b/man/setup.Rd
@@ -20,6 +20,9 @@ the model parameters.}
 \item{metrics}{(\code{list}, optional) A list of metrics to be tracked during
 the training procedure.}
 }
+\value{
+A luz module that can be trained with \code{\link[=fit]{fit()}}.
+}
 \description{
 The setup function is used to set important attributes and method for \code{nn_modules}
 to be used with Luz.
diff --git a/tests/testthat.R b/tests/testthat.R
index c1f0dd74..6bb2a3ff 100644
--- a/tests/testthat.R
+++ b/tests/testthat.R
@@ -1,4 +1,6 @@
 library(testthat)
 library(luz)
 
-test_check("luz")
+if (Sys.getenv("TORCH_TEST", unset = 0) == 1)
+  test_check("luz")
+