adjust_predicts in eval_method.py

[Zihan-Zhou]

This function currently adjusts predictions based on ground truth labels. This would increase the F1 score for example. However, during the inference stage, there won't be any labels available. Should predictions be adjusted by other method, or they remain as they are?

[lushangjun1998]

I also think this behavior is problematic, it may overestimate model performance.
The correct approach should be:

1. Remove function adjust_predicts, predictions should be derived solely from the threshold (score > threshold).
2. Add a post-processing operation if needed (e.g, smoothing).

Unfortunately, many subsequent projects based on this implementation have adopted the same approach, leading to unreliable evaluation results.