You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
您好,您的工作使人眼前一亮!
有个问题请教一下。对于小样本分割等任务,是使用 episode 的模式,对 query image 预测 mask,然后与 ground truth 计算Loss。
不太明白您这个工作中Loss是怎样计算的,为什么有 K 个 support view,就有 K 个 Loss呢,难道是与 support view 的ground truth 计算Loss?
The text was updated successfully, but these errors were encountered:
您好,您的工作使人眼前一亮!
有个问题请教一下。对于小样本分割等任务,是使用 episode 的模式,对 query image 预测 mask,然后与 ground truth 计算Loss。
不太明白您这个工作中Loss是怎样计算的,为什么有 K 个 support view,就有 K 个 Loss呢,难道是与 support view 的ground truth 计算Loss?
The text was updated successfully, but these errors were encountered: