DSPy: The framework for programming—not prompting—language models
Python
Fix evaluation multithreaded results
liorfo6fda4d2f8f39
diff --git a/dspy/evaluate/evaluate.py b/dspy/evaluate/evaluate.pyindex 603b552aa..8c0e609e0 100644--- a/dspy/evaluate/evaluate.py+++ b/dspy/evaluate/evaluate.py@@ -166,9 +166,6 @@ class Evaluate:num_threads,display_progress,)- if return_outputs: # Handle the return_outputs logic- results = [(example, prediction, score)- for _, example, prediction, score in reordered_devset]if display:print(@@ -176,6 +173,10 @@ class Evaluate:predicted_devset = sorted(reordered_devset)+ if return_outputs: # Handle the return_outputs logic+ results = [(example, prediction, score)+ for _, example, prediction, score in predicted_devset]+# data = [{**example, **prediction, 'correct': score} for example, prediction, score in zip(reordered_devset, preds, scores)]data = [merge_dicts(example, prediction) | {"correct": score} for _, example, prediction, score in predicted_devset@@ -222,9 +223,9 @@ class Evaluate:ipython_display(HTML(message))if return_all_scores and return_outputs:- return round(100 * ncorrect / ntotal, 2), results, [score for *_, score in reordered_devset]+ return round(100 * ncorrect / ntotal, 2), results, [score for *_, score in predicted_devset]elif return_all_scores:- return round(100 * ncorrect / ntotal, 2), [score for *_, score in reordered_devset]+ return round(100 * ncorrect / ntotal, 2), [score for *_, score in predicted_devset]elif return_outputs:return round(100 * ncorrect / ntotal, 2), results