BalancedMSE

BalancedMSE copied to clipboard

about GAILoss

1

Thank you for your work. gmm = {k: gmm[k].reshape(1, -1).expand(pred.shape[0], -1) for k in gmm} mse_term = F.mse_loss(pred, target, reduction='none') / 2 / noise_var + 0.5 * noise_var.log() sum_var =...

luxianlebron

About the "bmc_loss_md"

3

Thanks for your work. def bmc_loss_md(pred, target, noise_var): I = torch.eye(pred.shape[-1]) logits = MVN(pred.unsqueeze(1), noise_var*I).log_prob(target.unsqueeze(0)) loss = F.cross_entropy(logits, torch.arange(pred.shape[0])) loss = loss * (2 * noise_var).detach() return loss My size...

SongYxing

When using the Multi-dimensional Balanced MSE, there is a memory error

4

When using the Multi-dimensional Balanced MSE for pixel-wise loss，there is an error：not enough memory: you tried to allocate 9663676416 bytes. Buy new RAM!

bowen480

Negative loss while using BNI implementation option

1

Hi @jiawei-ren I am trying to implement the Balanced MSE to other long tail regression task. Firstly, I tried the BNI, and then the negative traning loss are often computed...

CeceliaSoh

Should one scale multi-dimensional targets when training with BMCLossMD?

1

Hello, thank you for this great work. I am currently trying to train a network for multi-task regression using BMCLossMD. However, the distributions of the target variables are different. For...

yanisnotavocado