diff --git a/R/LearnerTorch.R b/R/LearnerTorch.R
index 10466256..fcb9c942 100644
--- a/R/LearnerTorch.R
+++ b/R/LearnerTorch.R
@@ -452,15 +452,6 @@ LearnerTorch = R6Class("LearnerTorch",
       # Ideally we could rely on state$train_task, but there is this complication
       # https://github.com/mlr-org/mlr3/issues/947
       param_vals$device = auto_device(param_vals$device)
-      if (!test_equal_col_info(ci_train, ci_predict)) { # nolint
-        stopf(paste0(
-          "Predict task's column info does not match the train task's column info.\n",
-          "This migth be handled more gracefully in the future.\n",
-          "Training column info:\n'%s'\n",
-          "Prediction column info:\n'%s'"),
-          paste0(capture.output(ci_train), collapse = "\n"),
-          paste0(capture.output(ci_predict), collapse = "\n"))
-      }
       private$.verify_predict_task(task, param_vals)
 
       with_torch_settings(seed = self$model$seed, num_threads = param_vals$num_threads, {
diff --git a/tests/testthat/test_learner_torch_methods.R b/tests/testthat/test_learner_torch_methods.R
index 673b1e4e..6f1088df 100644
--- a/tests/testthat/test_learner_torch_methods.R
+++ b/tests/testthat/test_learner_torch_methods.R
@@ -178,5 +178,5 @@ test_that("wrong column info stops learner from prediction.", {
 
   learner = lrn("regr.torch_featureless", epochs = 1, batch_size = 50)
   learner$train(t1)
-  expect_error(learner$predict(t2), "more gracefully")
+  expect_error(learner$predict(t2))
 })