ginka-generator/ginka/vae_rnn/loss.py
2026-02-06 14:31:07 +08:00

18 lines
521 B
Python

import torch
import torch.nn.functional as F
class VAELoss:
def __init__(self):
self.num_classes = 32
def vae_loss(self, logits, target, mu, logvar, beta=0.1):
# target: [B, 13, 13]
target = F.one_hot(target, num_classes=self.num_classes).float().permute(0, 3, 1, 2)
recon_loss = F.cross_entropy(logits, target)
kl_loss = -0.5 * torch.mean(
1 + logvar - mu.pow(2) - logvar.exp()
)
return recon_loss + beta * kl_loss, recon_loss, kl_loss