I'm wondering if it is possible to extract the class confidence and location box loss for each image in the batch, so not having a unique loss metric at the end but a tensor of loss in the such form [batch_size, 2] with each column corresponding to the class confidence and location box loss?
I'm asking this since I'm willing to add an additional module up on SSD and I need the loss for each images in the batch.
I'm wondering if it is possible to extract the class confidence and location box loss for each image in the batch, so not having a unique loss metric at the end but a tensor of loss in the such form [batch_size, 2] with each column corresponding to the class confidence and location box loss?
I'm asking this since I'm willing to add an additional module up on SSD and I need the loss for each images in the batch.