训练结果

[AllesOderNicht]

在got10k，tiny训练时，使用RTX3090，BS(32)进行训练时，在30个epoch的时候已经达到了68的准确率，在训练就发生了过拟合，这正常吗。我看作者的论文是训练了300个epoch，210时学习率衰减。。。

[Xrenya]

@AllesOderNicht The paper states that they trained on 4 datasets, not separately.

Training. We train SwinTrack using training splits of LaSOT [15], TrackingNet [33], GOT-10k [20] (1,000 videos are removed for fair comparisons with other trackers [22, 44]) and COCO 2017 [29].

[sansanfree]

@AllesOderNicht The paper states that they trained on 4 datasets, not separately.

Training. We train SwinTrack using training splits of LaSOT [15], TrackingNet [33], GOT-10k [20] (1,000 videos are removed for fair comparisons with other trackers [22, 44]) and COCO 2017 [29].

GOT-10k [20] provides 180 sequences for testing and it requires trackers to be trained using GOT-10k train split only