mlr-org · sebffischer · Nov 21, 2025 · Nov 12, 2025
diff --git a/R/LearnerTorch.R b/R/LearnerTorch.R
@@ -44,6 +44,17 @@
 #' * multi-class classification: The `factor` target variable of a [`TaskClassif`][mlr3::TaskClassif] is a label-encoded
 #'   [`torch_long`][torch::torch_long] with shape `(batch_size)` where the label-encoding goes from `1` to `n_classes`.
 #'
+#' @section Important Runtime Considerations:
+#' There are a few hyperparameters settings that can have a considerable impact on the runtime of the learner.
+#' These include:
+#'
+#' * `device`: Use a GPU if possible.
+#' * `num_threads`: Set this to the number of CPU cores available if training on CPU.
+#' * `tensor_dataset`: Set this to `TRUE` (or `"device"` if on a GPU) if the dataset fits into memory.
+#' * `batch_size`: Especially for very small models, choose a larger batch size.
+#'
+#' Also, see the *Early Stopping and Internal Tuning* section for how to terminate training early.
+#'
 #' @template param_id
 #' @template param_task_type
 #' @template param_param_vals

diff --git a/man-roxygen/paramset_torchlearner.R b/man-roxygen/paramset_torchlearner.R
@@ -62,7 +62,6 @@
 #' **Dataloader**:
 #' * `batch_size` :: `integer(1)`\cr
 #'   The batch size (required).
-#'   When working with small models or datasets, choosing a larger batch size can considerably speed up training.
 #' * `shuffle` :: `logical(1)`\cr
 #'   Whether to shuffle the instances in the dataset. This is initialized to `TRUE`,
 #'   which differs from the default (`FALSE`).

diff --git a/man/mlr_learners_torch.Rd b/man/mlr_learners_torch.Rd
diff --git a/man/mlr_pipeops_torch_model.Rd b/man/mlr_pipeops_torch_model.Rd