start saving version of library in the checkpoints, so in case of breaking changes, researchers can return to the right working version

denoising_diffusion_pytorch/denoising_diffusion_pytorch.py

@@ -431,7 +431,7 @@ class GaussianDiffusion(nn.Module):
         beta_schedule = 'cosine',
         p2_loss_weight_gamma = 0., # p2 loss weight, from https://arxiv.org/abs/2204.00227 - 0 is equivalent to weight of 1 across time - 1. is recommended
         p2_loss_weight_k = 1,
-        ddim_sampling_eta = 0.
+        ddim_sampling_eta = 1.
     ):
         super().__init__()
         assert not (type(self) == GaussianDiffusion and model.channels != model.out_dim)
@@ -810,8 +810,8 @@ class Trainer(object):
         if self.accelerator.is_main_process:
             self.ema = EMA(diffusion_model, beta = ema_decay, update_every = ema_update_every)
 
-        self.results_folder = Path(results_folder)
-        self.results_folder.mkdir(exist_ok = True)
+            self.results_folder = Path(results_folder)
+            self.results_folder.mkdir(exist_ok = True)
 
         # step counter state
 

denoising_diffusion_pytorch/denoising_diffusion_pytorch_1d.py

@@ -414,7 +414,7 @@ class GaussianDiffusion1D(nn.Module):
         beta_schedule = 'cosine',
         p2_loss_weight_gamma = 0.,
         p2_loss_weight_k = 1,
-        ddim_sampling_eta = 0.
+        ddim_sampling_eta = 1.
     ):
         super().__init__()
         self.model = model

denoising_diffusion_pytorch/version.py

@@ -1 +1 @@
-__version__ = '0.1.5'
+__version__ = '1.16.6'