Closed hahapt closed 1 year ago
We did a grid search on the score threshold and update_score_threshold during inference.
A corresponding threshold in training is also set to 0.5. https://github.com/megvii-research/MOTRv2/blob/be49b7336218e470c9ebcd34be54fe7eec702675/models/qim.py#L70
I found the score threshold and update_score_threshold is set to 0.5 in submit_dance.py, then the detections under 0.5 confidence will directly be removed.
So why it set to be 0.5? Is there any hyper-parameter during training influence them?