diff --git a/R/sits_classify.R b/R/sits_classify.R
index 725aa949d..cbf177821 100644
--- a/R/sits_classify.R
+++ b/R/sits_classify.R
@@ -39,6 +39,7 @@
 #' @param  multicores        Number of cores to be used for classification
 #'                           (integer, min = 1, max = 2048).
 #' @param  gpu_memory        Memory available in GPU in GB (default = 4)
+#' @param  batch_size        Batch size for GPU classification.
 #' @param  n_sam_pol         Number of time series per segment to be classified
 #'                           (integer, min = 10, max = 50).
 #' @param  output_dir        Valid directory for output file.
@@ -79,7 +80,11 @@
 #'    Use an sf object or a shapefile to define it.
 #'
 #'    When using a GPU for deep learning, \code{gpu_memory} indicates the
-#'    memory of available in the graphics card.
+#'    memory of available in the graphics card. The parameter \code{batch_size}
+#'    defines the size of the matrix (measured in number of rows)
+#'    which is sent to the GPU for classification. Users can test
+#'    different sizes to best fit their GPU architecture.
+#'
 #'    It is not possible to have an exact idea of the size of Deep Learning
 #'    models in GPU memory, as the complexity of the model and factors
 #'    such as CUDA Context increase the size of the model in memory.
diff --git a/R/sits_lighttae.R b/R/sits_lighttae.R
index 1760db0ad..faa3938d2 100644
--- a/R/sits_lighttae.R
+++ b/R/sits_lighttae.R
@@ -310,14 +310,14 @@ sits_lighttae <- function(samples = NULL,
             values <- array(
                 data = as.matrix(values), dim = c(n_samples, n_times, n_bands)
             )
-            # Get batch size
-            batch_size <- sits_env[["batch_size"]]
             # if CUDA is available and gpu memory is defined, transform values
             # to torch dataloader
-            if (.torch_has_cuda() && .has(gpu_memory)) {
+            if (.torch_has_cuda()) {
+                # Get batch size
+                batch_size <- sits_env[["batch_size"]]
                 # transfor the input array to a dataset
                 values <- .as_dataset(values)
-                # To the data set to a torcj  transform in a dataloader to use the batch size
+                # To the data set to a torch  transform in a dataloader to use the batch size
                 values <- torch::dataloader(values, batch_size = batch_size)
                 # Do GPU classification with dataloader
                 values <- .try(
diff --git a/R/sits_mlp.R b/R/sits_mlp.R
index 945643a43..50df44452 100644
--- a/R/sits_mlp.R
+++ b/R/sits_mlp.R
@@ -250,13 +250,11 @@ sits_mlp <- function(samples = NULL,
             values <- .pred_normalize(pred = values, stats = ml_stats)
             # Transform input into matrix
             values <- as.matrix(values)
-            # Get batch size
-            batch_size <- sits_env[["batch_size"]]
             # if CUDA is available and gpu memory is defined, transform values
             # to torch dataloader
-            if (.torch_has_cuda() && .has(gpu_memory)) {
-                # set the batch size according to the GPU memory
-                b_size <- 2^gpu_memory
+            if (.torch_has_cuda()) {
+                # Get batch size
+                batch_size <- sits_env[["batch_size"]]
                 # transfor the input array to a dataset
                 values <- .as_dataset(values)
                 # To the data set to a torcj  transform in a dataloader to use the batch size
diff --git a/R/sits_tae.R b/R/sits_tae.R
index f68286a5b..829f05383 100644
--- a/R/sits_tae.R
+++ b/R/sits_tae.R
@@ -277,12 +277,12 @@ sits_tae <- function(samples = NULL,
             values <- array(
                 data = as.matrix(values), dim = c(n_samples, n_times, n_bands)
             )
-            # Get batch size
-            batch_size <- sits_env[["batch_size"]]
             # if CUDA is available and gpu memory is defined, transform values
             # to torch dataloader
-            if (.torch_has_cuda() && .has(gpu_memory)) {
-                # transfor the input array to a dataset
+            if (.torch_has_cuda()) {
+                # Get batch size
+                batch_size <- sits_env[["batch_size"]]
+                # transform the input array to a dataset
                 values <- .as_dataset(values)
                 # To the data set to a torcj  transform in a dataloader to use the batch size
                 values <- torch::dataloader(values, batch_size = batch_size)
diff --git a/R/sits_tempcnn.R b/R/sits_tempcnn.R
index cc5b03dd3..4ba2c3751 100644
--- a/R/sits_tempcnn.R
+++ b/R/sits_tempcnn.R
@@ -323,12 +323,12 @@ sits_tempcnn <- function(samples = NULL,
             values <- array(
                 data = as.matrix(values), dim = c(n_samples, n_times, n_bands)
             )
-            # Get batch size
-            batch_size <- sits_env[["batch_size"]]
             # if CUDA is available and gpu memory is defined, transform values
             # to torch dataloader
             if (.torch_has_cuda()) {
-                # transfor the input array to a dataset
+                # Get batch size
+                batch_size <- sits_env[["batch_size"]]
+                # transform the input array to a dataset
                 values <- .as_dataset(values)
                 # To the data set to a torch transform in a dataloader to use the batch size
                 values <- torch::dataloader(values, batch_size = batch_size)