For the Evaluation section, which files do ‘PATH_TO_LABEL_JSON_FILE}’ and ‘{PATH_TO_PREDICTION_JSON_FILE}’ specifically refer to.

[27182812]

python ./eval/eval_task.py
--golds_json {PATH_TO_LABEL_JSON_FILE}
--preds_json {PATH_TO_PREDICTION_JSON_FILE}
--task_name {TASK_ID}
--output_file {RESULT_JSON_PATH}

Good work, I am reproducing this work. I have a question.
There are many stages. In the Evaluation section, which files do ‘PATH_TO_LABEL_JSON_FILE}’ and ‘{PATH_TO_PREDICTION_JSON_FILE}’ specifically refer to?

[Freedom2147483647]

@27182812
Excuse me. What were the results you reproduced after running the task_LoRA.py?
I couldn't fully reproduce the data results in the original paper. My results are shown in the figure below. The parts marked in yellow are the ones that are higher than the data in the original paper.

[Freedom2147483647]

In the labeled cells, the black part on the left is the result reproduced by me, and the red part on the right is the data in the original paper.