Improve test (#674)

The idea is to read the expected results before running the model so that if they are not fetched with Git LFS, the user receives an error before running the model on the data.

Author: qubvel

tests/slow_tests/test_accelerate_model.py

@@ -86,18 +86,20 @@ def test_accelerate_model_prediction(tests: list[ModelInput]):
     """Evaluates a model on a full task - is parametrized using pytest_generate_test"""
     model_args, get_predictions = tests
 
-    predictions = get_predictions()["results"]
-
     # Load the reference results
     with open(model_args["results_file"], "r") as f:
         reference_results = json.load(f)["results"]
 
     # Change the key names, replace '|' with ':'
     reference_results = {k.replace("|", ":"): v for k, v in reference_results.items()}
 
+    # Get the predictions
+    predictions = get_predictions()["results"]
+
     # Convert defaultdict values to regular dict for comparison
     predictions_dict = {k: dict(v) if hasattr(v, "default_factory") else v for k, v in predictions.items()}
 
+    # Compare the predictions with the reference results
     diff = DeepDiff(reference_results, predictions_dict, ignore_numeric_type_changes=True, math_epsilon=0.05)
 
     assert diff == {}, f"Differences found: {diff}"