diff --git a/examples/legacy/seq2seq/finetune_trainer.py b/examples/legacy/seq2seq/finetune_trainer.py
index e9daf9fc506d..44f5a75eda0a 100755
--- a/examples/legacy/seq2seq/finetune_trainer.py
+++ b/examples/legacy/seq2seq/finetune_trainer.py
@@ -32,7 +32,7 @@
     MBartTokenizerFast,
     set_seed,
 )
-from transformers.trainer_utils import EvaluationStrategy, is_main_process
+from transformers.trainer_utils import is_main_process
 from transformers.training_args import ParallelMode
 from utils import (
     Seq2SeqDataCollator,
@@ -271,7 +271,7 @@ def main():
             max_source_length=data_args.max_source_length,
             prefix=model.config.prefix or "",
         )
-        if training_args.do_eval or training_args.eval_strategy != EvaluationStrategy.NO
+        if training_args.do_eval
         else None
     )
     test_dataset = (
diff --git a/src/transformers/trainer_utils.py b/src/transformers/trainer_utils.py
index 951e7b52017d..331b0b61d74f 100644
--- a/src/transformers/trainer_utils.py
+++ b/src/transformers/trainer_utils.py
@@ -224,12 +224,6 @@ class SaveStrategy(ExplicitEnum):
     BEST = "best"
 
 
-class EvaluationStrategy(ExplicitEnum):
-    NO = "no"
-    STEPS = "steps"
-    EPOCH = "epoch"
-
-
 class HubStrategy(ExplicitEnum):
     END = "end"
     EVERY_SAVE = "every_save"
diff --git a/src/transformers/training_args.py b/src/transformers/training_args.py
index 3520ea5702aa..ca2bc0dcacea 100644
--- a/src/transformers/training_args.py
+++ b/src/transformers/training_args.py
@@ -25,7 +25,6 @@
 
 from .debug_utils import DebugOption
 from .trainer_utils import (
-    EvaluationStrategy,
     FSDPOption,
     HubStrategy,
     IntervalStrategy,
@@ -1482,15 +1481,6 @@ def __post_init__(self):
         if self.disable_tqdm is None:
             self.disable_tqdm = logger.getEffectiveLevel() > logging.WARN
 
-        if isinstance(self.eval_strategy, EvaluationStrategy):
-            warnings.warn(
-                "using `EvaluationStrategy` for `eval_strategy` is deprecated and will be removed in version 5"
-                " of 🤗 Transformers. Use `IntervalStrategy` instead",
-                FutureWarning,
-            )
-            # Go back to the underlying string or we won't be able to instantiate `IntervalStrategy` on it.
-            self.eval_strategy = self.eval_strategy.value
-
         self.eval_strategy = IntervalStrategy(self.eval_strategy)
         self.logging_strategy = IntervalStrategy(self.logging_strategy)
         self.save_strategy = SaveStrategy(self.save_strategy)