Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix output format of airbench and vocal sound #510

Merged
merged 1 commit into from
Feb 1, 2025
Merged

fix output format of airbench and vocal sound #510

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 17 additions & 3 deletions lmms_eval/tasks/air_bench/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -132,7 +132,10 @@ def air_bench_process_results_chat(doc, result):
eval_answer2, model_name2 = get_eval(max_tokens=1024, content=content)

return {
"gpt_eval": {"eval_answer": [eval_answer, eval_answer2], "model_name": model_name},
"gpt_eval": {
"eval_answer": [eval_answer, eval_answer2],
"model_name": model_name,
},
}


Expand Down Expand Up @@ -189,7 +192,10 @@ def air_bench_process_results_foundation(doc, result):
score = 1.0 if pred == gt_ans else 0.0
submission_dict = {}
submission_dict = {doc.get("uniq_id", "unknown"): pred}
return {"accuracy": {"score": score, "task": doc["task_name"]}, "submission": submission_dict}
return {
"accuracy": {"score": score, "task": doc["task_name"]},
"submission": submission_dict,
}


def air_bench_aggregate_results_for_submission(results, args):
Expand All @@ -211,7 +217,15 @@ def air_bench_aggregate_results_foundation(results):
categorical_correct[result["task"]] += result["score"]
categorical_total[result["task"]] += 1

return {"overall_accuracy": score / len(results), "categorical_accuracy": {task: categorical_correct[task] / categorical_total[task] for task in categorical_correct.keys()}}
overall_accuracy = round(score / len(results), 5)
categorical_accuracy = {task: round(categorical_correct[task] / categorical_total[task], 5) for task in categorical_correct.keys()}

eval_logger.info("=" * 50)
eval_logger.info(f"Overall accuracy: {overall_accuracy}")
for task, acc in categorical_accuracy.items():
eval_logger.info(f"{task} accuracy: {acc}")
eval_logger.info("=" * 50)
return overall_accuracy


def parse_multi_choice_response(response, all_choices):
Expand Down
23 changes: 14 additions & 9 deletions lmms_eval/tasks/vocalsound/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,16 +63,21 @@ def vocalsound_aggregate_results(results):
group_totals[age_group] += 1
group_correct[age_group] += accuracy

return {
"overall_accuracy": total_correct / len(results),
"categorical_accuracy": {
"male_accuracy": round(group_correct["male"] / group_totals.get("male", 1), 5), # Avoid division by zero
"female_accuracy": round(group_correct["female"] / group_totals.get("female", 1), 5),
"age_18_25_accuracy": round(group_correct["age1"] / group_totals.get("age1", 1), 5),
"age_26_48_accuracy": round(group_correct["age2"] / group_totals.get("age2", 1), 5),
"age_49_80_accuracy": round(group_correct["age3"] / group_totals.get("age3", 1), 5),
},
overall_accuracy = round(total_correct / len(results), 5)
categorical_accuracy = {
"male_accuracy": round(group_correct["male"] / group_totals.get("male", 1), 5), # Avoid division by zero
"female_accuracy": round(group_correct["female"] / group_totals.get("female", 1), 5),
"age_18_25_accuracy": round(group_correct["age1"] / group_totals.get("age1", 1), 5),
"age_26_48_accuracy": round(group_correct["age2"] / group_totals.get("age2", 1), 5),
"age_49_80_accuracy": round(group_correct["age3"] / group_totals.get("age3", 1), 5),
}
eval_logger.info("=" * 50)
eval_logger.info(f"Overall accuracy: {overall_accuracy}")
eval_logger.info("Categorical accuracy: ")
for key, value in categorical_accuracy.items():
eval_logger.info(f"{key} accuracy: {value}")
eval_logger.info("=" * 50)
return overall_accuracy


def get_answer(response):
Expand Down
Loading