Your browser does not seem to support JavaScript. As a result, your viewing experience will be diminished, and you have been placed in read-only mode.
Please download a browser that supports JavaScript, or enable it if it's disabled (i.e. NoScript).
我現在自行搭建了一個DETR模型，它在單張圖像上的表現還不錯，但如果是訓練兩張圖片時，需要更長的時間去收斂，並且兩張物件的bbox很容易互相出現，請問有甚麼解決辦法或是其他有關DETR訓練需要注意的事項。
