Fit test for ensembles

Author: han-ol

tests/test_approximators/test_approximator_ensemble/test_fit.py

@@ -0,0 +1,35 @@
+import re
+
+import io
+from contextlib import redirect_stdout
+
+
+def test_loss_progress(continuous_approximator_ensemble, train_dataset_for_ensemble, validation_dataset):
+    continuous_approximator_ensemble.compile(optimizer="AdamW")
+    num_epochs = 3
+
+    # Capture ostream and train model
+    with io.StringIO() as stream:
+        with redirect_stdout(stream):
+            continuous_approximator_ensemble.fit(
+                dataset=train_dataset_for_ensemble, validation_data=validation_dataset, epochs=num_epochs
+            )
+
+        output = stream.getvalue()
+
+    print(output)
+
+    # check that there is a progress bar
+    assert "━" in output, "no progress bar"
+
+    # check that the loss is shown
+    assert "loss" in output
+    assert re.search(r"\bloss: \d+\.\d+", output) is not None, "training loss not correctly shown"
+
+    # check that validation loss is shown
+    assert "val_loss" in output
+    assert re.search(r"\bval_loss: \d+\.\d+", output) is not None, "validation loss not correctly shown"
+
+    # check that the shown loss is not nan or zero
+    assert re.search(r"\bnan\b", output) is None, "found nan in output"
+    assert re.search(r"\bloss: 0\.0000e\+00\b", output) is None, "found zero loss in output"