diff --git a/tests/conftest.py b/tests/conftest.py
index b6e9e68c..38ccaab2 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -65,5 +65,5 @@ def small_hcs_dataset(tmp_path_factory: TempPathFactory) -> Path:
 def labels_hcs_dataset(tmp_path_factory: TempPathFactory) -> Path:
     """Provides a small, not preprocessed HCS OME-Zarr dataset."""
     dataset_path = tmp_path_factory.mktemp("labels.zarr")
-    _build_hcs(dataset_path, ["DAPI"], (2, 16, 16), np.uint16, 3)
+    _build_hcs(dataset_path, ["DAPI", "GFP"], (2, 16, 16), np.uint16, 3)
     return dataset_path
diff --git a/tests/translation/test_evaluation.py b/tests/translation/test_evaluation.py
new file mode 100644
index 00000000..40cd8429
--- /dev/null
+++ b/tests/translation/test_evaluation.py
@@ -0,0 +1,33 @@
+import numpy as np
+import pandas as pd
+import pytest
+from lightning.pytorch.loggers import CSVLogger
+from numpy.testing import assert_array_equal
+
+from viscy.data.segmentation import SegmentationDataModule
+from viscy.trainer import Trainer
+from viscy.translation.evaluation import SegmentationMetrics2D
+
+
+@pytest.mark.parametrize("pred_channel", ["DAPI", "GFP"])
+def test_segmentation_metrics_2d(pred_channel, labels_hcs_dataset, tmp_path) -> None:
+    dm = SegmentationDataModule(
+        pred_dataset=labels_hcs_dataset,
+        target_dataset=labels_hcs_dataset,
+        target_channel="DAPI",
+        pred_channel=pred_channel,
+        pred_z_slice=0,
+        target_z_slice=0,
+        batch_size=1,
+        num_workers=0,
+    )
+    lm = SegmentationMetrics2D()
+    trainer = Trainer(logger=CSVLogger(tmp_path, name="", version=""))
+    trainer.test(lm, datamodule=dm)
+    metrics = pd.read_csv(tmp_path / "metrics.csv")
+    if pred_channel == "DAPI":
+        assert_array_equal(
+            metrics["accuracy"].to_numpy(), np.ones_like(metrics["accuracy"])
+        )
+    else:
+        assert 0 < metrics["accuracy"].mean() < 1