HI, Currently, I use KL criterion for my loss function. I do not understand what means in the instruction by" By default, the losses are averaged for each minibatch over observations as well as over dimensions. However, if the field sizeAverage is set to false, the losses are instead summed.
" see(https://github.com/torch/nn/blob/master/doc/criterion.md)
Does it means the KL also the average of the label dimension? for example, if a=[0.1, 0.2, 0.9;0.2,0.3,0.5] b =[0.2,0.4,0.4;0.5,0.2,0.3]
The KL loss will division 2(for the batch size) or division( 2*3(label dimensional))?
HI, Currently, I use KL criterion for my loss function. I do not understand what means in the instruction by" By default, the losses are averaged for each minibatch over observations as well as over dimensions. However, if the field sizeAverage is set to false, the losses are instead summed. " see(https://github.com/torch/nn/blob/master/doc/criterion.md) Does it means the KL also the average of the label dimension? for example, if a=[0.1, 0.2, 0.9;0.2,0.3,0.5] b =[0.2,0.4,0.4;0.5,0.2,0.3] The KL loss will division 2(for the batch size) or division( 2*3(label dimensional))?