zzzmm1

Hi, thanks for your good work! Could you please provide demo code for the experiments in Table 1, e.g the AP 74.7 using gt_cls_iou?

resume的时候会遇到当时训练的第一个epoch的lr在log中显示是0.01，而下一个epoch又恢复到正常的那个epoch该有的lr，这是一个bug吗？

i meet the same question, there is motion_change.label in my downloaded VOT2019 file. how do u solve it?

RuntimeError: The size of tensor a (10000) must match the size of tensor b (4624) at non-singleton dimension 0 similar problem

When i trained siammask_sharp, i met the same problem.

> @qxzhou1010 You can also change line https://github.com/nmslib/hnswlib/blob/master/hnswlib/hnswalg.h#L556 From `searchBaseLayerST` to `searchBaseLayerST`. It will start collecting the number of calculations (`metric_distance_computations`, `metric_hops` fields) until you manually reset them. To access...

Hello @searchivarius If we use data slices to search a query in parallel, the distance required to search the graph constructed by each data slice is roughly the same as...

作者您好，请问您使用多GPU训练吗？有这部分代码吗？

Thanks very much for your reply. I understand your preprocessing method, and I got a copy of train_data.json and test_data.json, but I don’t know if they are the same as...