fix few Help typos (#147)

* fix help typo * remove duplicate arg * fix typo * fix typo in csv_logger * fix typo * fix typo. * fix typos in callback * fix typo * add new author
mlverse · Jul 30, 2024 · cafc744 · cafc744
1 parent fdc660b
commit cafc744
Show file tree

Hide file tree

Showing 14 changed files with 30 additions and 34 deletions.
diff --git a/DESCRIPTION b/DESCRIPTION
@@ -3,6 +3,7 @@ Title: Higher Level 'API' for 'torch'
 Version: 0.4.0.9002
 Authors@R: c(
     person("Daniel", "Falbel", email = "[email protected]", role = c("aut", "cre", "cph")),
+    person("Christophe", "Regouby", email = "[email protected]", role = c("ctb")),
     person(family = "RStudio", role = c("cph"))
     )
 Description: A high level interface for 'torch' providing utilities to reduce the

diff --git a/R/callbacks-resume.R b/R/callbacks-resume.R
@@ -131,7 +131,7 @@ luz_callback_auto_resume <- luz_callback(
 #' Allow resume model training from a specific checkpoint
 #'
 #' @param path Path to the checkpoint that you want to resume.
-#' @param restore_model_state Wether to restore the model state from the callback.
+#' @param restore_model_state Wether to restore the model state from the checkpoint.
 #' @param restore_records Wether to restore records from the checkpoint.
 #' @param restore_optimizer_state Wether to restore the optimizer state from the
 #'   checkpoint.

diff --git a/R/callbacks.R b/R/callbacks.R
@@ -84,7 +84,7 @@ default_evaluate_callbacks <- function() {
 #' @section Prediction callbacks:
 #'
 #' You can also use callbacks when using [predict()]. In this case the supported
-#' callback methods are detailed above.
+#' callback methods are detailed below:
 #'
 #' ```
 #' Start predict
@@ -382,7 +382,7 @@ luz_callback_metrics <- luz_callback(
 #' - `ctx$loss`: Resets the `loss` attribute to `list()` when finished training/ or
 #'   validating.
 #'
-#' @note In general you won't need to explicitly use the metrics callback as it's
+#' @note In general you won't need to explicitly use the train_valid callback as it's
 #' used by default in [fit.luz_module_generator()].
 #'
 #' @returns
@@ -466,7 +466,7 @@ luz_callback_lr_scheduler <- luz_callback(
 
 #' CSV logger callback
 #'
-#' Logs metrics obtained during training a fiel on disk.
+#' Logs metrics obtained during training a file on disk.
 #' The file will have 1 line for each epoch/validation.
 #'
 #' @param path path to a file on disk.

diff --git a/R/context.R b/R/context.R
@@ -17,10 +17,9 @@ NULL
 #' See also [ctx].
 #'
 #' @param name name of the metric
-#' @param value value to log
 #' @param what (string) What you are logging.
 #' @param set (string) Usually 'train' or 'valid' indicating the set you want
-#'  to lot to. But can be arbitrary info.
+#'  to log to. But can be arbitrary info.
 #' @param value Arbitrary value to log.
 #' @param index Index that this value should be logged. If `NULL` the value
 #'  is added to the end of list, otherwise the index is used.
@@ -80,7 +79,7 @@ context <- R6::R6Class(
       invisible(self)
     },
     #' @description
-    #' Log a metric gen its name and value.
+    #' Log a metric by its name and value.
     #' Metric values are indexed by epoch.
     log_metric = function(name, value) {
       set <- if (self$training) "train" else "valid"

diff --git a/R/serialization.R b/R/serialization.R
@@ -13,7 +13,7 @@
 #'
 #' @param obj an object of class 'luz_module_fitted' as returned by
 #' [fit.luz_module_generator()].
-#' @param path path in file system so save the object.
+#' @param path path in file system to the object.
 #' @param ... currently unused.
 #'
 #' @family luz_save
@@ -137,7 +137,7 @@ luz_checkpoint <- function(ctx, path) {
 #'
 #' Works with checkpoints created typically with [luz_callback_model_checkpoint()].
 #'
-#' @param obj Object to which we want to laod the checkpoint.
+#' @param obj Object to which we want to load the checkpoint.
 #' @param path Path of the checkpoint on disk.
 #' @param ... unused. Is there to allow future extensions.
 #'

diff --git a/man/context.Rd b/man/context.Rd
diff --git a/man/luz_callback.Rd b/man/luz_callback.Rd
diff --git a/man/luz_callback_csv_logger.Rd b/man/luz_callback_csv_logger.Rd
diff --git a/man/luz_callback_resume_from_checkpoint.Rd b/man/luz_callback_resume_from_checkpoint.Rd
diff --git a/man/luz_callback_train_valid.Rd b/man/luz_callback_train_valid.Rd
diff --git a/man/luz_load.Rd b/man/luz_load.Rd
diff --git a/man/luz_load_checkpoint.Rd b/man/luz_load_checkpoint.Rd
diff --git a/man/luz_save.Rd b/man/luz_save.Rd
diff --git a/man/rmd/callbacks.Rmd b/man/rmd/callbacks.Rmd
@@ -73,7 +73,7 @@ Here's an overview of possible callback *breakpoints*:
        - on_fit_end
     End Fit
 
-Every step market with `on_*` is a point in the training procedure that is available for callbacks to be called.
+Every step marked with `on_*` is a point in the training procedure that is available for callbacks to be called.
 
 The other important part of callbacks is the `ctx` (context) object. 
 See `help("ctx")` for details.