Open-Sora: Democratizing Efficient Video Production for All
Python
debug
Shen-Chenhui735e9e722b8f
diff --git a/opensora/models/vae/vae_3d_v2.py b/opensora/models/vae/vae_3d_v2.pyindex ff376a1..8a81e2c 100644--- a/opensora/models/vae/vae_3d_v2.py+++ b/opensora/models/vae/vae_3d_v2.py@@ -691,7 +691,7 @@ class VAE_3D_V2(nn.Module):# # Perceptual Loss# self.vgg = None# self.perceptual_loss_weight = perceptual_loss_weight- # if perceptual_loss_weight is not None and perceptual_loss_weight > 0:+ # if perceptual_loss_weight is not None and perceptual_loss_weight > 0.0:# # self.lpips = LPIPS().eval()# if not exists(vgg):# vgg = torchvision.models.vgg16(@@ -704,7 +704,7 @@ class VAE_3D_V2(nn.Module):# self.discriminator_factor = discriminator_factor# self.discriminator_start = discriminator_start# self.discriminator = None- # if discriminator_factor is not None and discriminator_factor > 0:+ # if discriminator_factor is not None and discriminator_factor > 0.0:# self.discriminator = StyleGANDiscriminator(# image_size = image_size,# num_frames = num_frames,@@ -842,7 +842,7 @@ class VAE_3D_V2(nn.Module):# # KL Loss# weighted_kl_loss = 0- # if self.kl_loss_weight is not None and self.kl_loss_weight > 0:+ # if self.kl_loss_weight is not None and self.kl_loss_weight > 0.0:# kl_loss = posterior.kl()# # NOTE: since we use MSE, here use mean as well, else use sum# kl_loss = torch.mean(kl_loss) / kl_loss.shape[0]@@ -852,7 +852,7 @@ class VAE_3D_V2(nn.Module):# # Perceptual Loss# # SCH: NOTE: if mse can pick single frame, if use sum of errors, need to calc for all frames!# weighted_perceptual_loss = 0- # if self.perceptual_loss_weight is not None and self.perceptual_loss_weight > 0:+ # if self.perceptual_loss_weight is not None and self.perceptual_loss_weight > 0.0:# frame_indices = torch.randn((batch, frames)).topk(1, dim = -1).indices# input_vgg_input = pick_video_frame(video, frame_indices)# recon_vgg_input = pick_video_frame(recon_video, frame_indices)@@ -954,7 +954,7 @@ class VEALoss(nn.Module):# KL Lossweighted_kl_loss = 0- if self.kl_loss_weight is not None and self.kl_loss_weight > 0:+ if self.kl_loss_weight is not None and self.kl_loss_weight > 0.0:kl_loss = posterior.kl()# NOTE: since we use MSE, here use mean as well, else use sumkl_loss = torch.mean(kl_loss) / kl_loss.shape[0]@@ -963,7 +963,7 @@ class VEALoss(nn.Module):# perceptual lossweighted_perceptual_loss = 0- if self.perceptual_loss_weight is not None and self.perceptual_loss_weight > 0:+ if self.perceptual_loss_weight is not None and self.perceptual_loss_weight > 0.0:assert exists(self.vgg)batch, channels, frames = video.shape[:3]frame_indices = torch.randn((batch, frames)).topk(1, dim = -1).indices