diff --git a/tests/test_sft_trainer.py b/tests/test_sft_trainer.py
index 9df7eeee8..eb2ca855a 100644
--- a/tests/test_sft_trainer.py
+++ b/tests/test_sft_trainer.py
@@ -164,9 +164,6 @@ def test_parse_arguments_peft_method(job_config):
 ############################# Prompt Tuning Tests #############################
 
 
-@pytest.mark.skip(
-    reason="currently inference doesn't work with transformer version 4.42.4"
-)
 def test_run_causallm_pt_and_inference():
     """Check if we can bootstrap and peft tune causallm models"""
     with tempfile.TemporaryDirectory() as tempdir:
@@ -197,9 +194,6 @@ def test_run_causallm_pt_and_inference():
         assert "### Text: @NortonSupport Thanks much.\n\n### Label:" in output_inference
 
 
-@pytest.mark.skip(
-    reason="currently inference doesn't work with transformer version 4.42.4"
-)
 def test_run_causallm_pt_and_inference_with_formatting_data():
     """Check if we can bootstrap and peft tune causallm models
     This test needs the trainer to format data to a single sequence internally.
@@ -238,9 +232,6 @@ def test_run_causallm_pt_and_inference_with_formatting_data():
         assert "### Text: @NortonSupport Thanks much.\n\n### Label:" in output_inference
 
 
-@pytest.mark.skip(
-    reason="currently inference doesn't work with transformer version 4.42.4"
-)
 def test_run_causallm_pt_and_inference_JSON_file_formatter():
     """Check if we can bootstrap and peft tune causallm models with JSON train file format"""
     with tempfile.TemporaryDirectory() as tempdir: