logs.txt

/home/mahesh/harsh/stable-diffusion/training.py:84: FutureWarning: `torch.cuda.amp.GradScaler(args...)` is deprecated. Please use `torch.amp.GradScaler('cuda', args...)` instead.
  self.scaler = torch.cuda.amp.GradScaler()
----------------------------------------------------------------------------------------------------
Loading pretrained models...
Models loaded successfully.
----------------------------------------------------------------------------------------------------
Creating dataloader...
Dataset vitonhd loaded, total 11647 pairs.
Training for 50 epochs
Batches per epoch: 5824
----------------------------------------------------------------------------------------------------
Initializing trainer...
Enabling PEFT training (self-attention layers only)
Total parameters: 899,226,667
Trainable parameters: 49,574,080 (5.51%)
Checkpoint loaded: ./checkpoints/checkpoint_step_50000.pth
Resuming from epoch 13, step 50000
Starting training...
Starting training for 50 epochs
Total training batches per epoch: 5824
Using DREAM with lambda = 0
Mixed precision: True
/home/mahesh/harsh/stable-diffusion/training.py:304: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast():
/home/mahesh/harsh/stable-diffusion/training.py:194: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(enabled=self.use_mixed_precision):
/home/mahesh/harsh/stable-diffusion/utils.py:491: FutureWarning: `torch.cuda.amp.autocast(args...)` is deprecated. Please use `torch.amp.autocast('cuda', args...)` instead.
  with torch.cuda.amp.autocast(enabled=False):