Add magic to debug

invoke-ai · Aug 14, 2023 · 511da59 · 511da59
1 parent 409e5d0
commit 511da59
Showing 1 changed file with 20 additions and 15 deletions.
diff --git a/invokeai/backend/stable_diffusion/diffusers_pipeline.py b/invokeai/backend/stable_diffusion/diffusers_pipeline.py
@@ -92,7 +92,7 @@ class AddsMaskGuidance:
     mask: torch.FloatTensor
     mask_latents: torch.FloatTensor
     scheduler: SchedulerMixin
-    noise: torch.Tensor
+    noise: Optional[torch.Tensor]
 
     def __call__(self, step_output: Union[BaseOutput, SchedulerOutput], t: torch.Tensor, conditioning) -> BaseOutput:
         output_class = step_output.__class__  # We'll create a new one with masked data.
@@ -124,7 +124,10 @@ def apply_mask(self, latents: torch.Tensor, t) -> torch.Tensor:
             t = einops.repeat(t, "-> batch", batch=batch_size)
         # Noise shouldn't be re-randomized between steps here. The multistep schedulers
         # get very confused about what is happening from step to step when we do that.
-        mask_latents = self.scheduler.add_noise(self.mask_latents, self.noise, t)
+        if self.noise is not None:
+            mask_latents = self.scheduler.add_noise(self.mask_latents, self.noise, t)
+        else:
+            mask_latents = self.mask_latents.clone()
         # TODO: Do we need to also apply scheduler.scale_model_input? Or is add_noise appropriately scaled already?
         # mask_latents = self.scheduler.scale_model_input(mask_latents, t)
         mask_latents = einops.repeat(mask_latents, "b c h w -> (repeat b) c h w", repeat=batch_size)
@@ -368,19 +371,21 @@ def latents_from_embeddings(
                 # TODO: we should probably pass this in so we don't have to try/finally around setting it.
                 self.invokeai_diffuser.model_forward_callback = AddsMaskLatents(self._unet_forward, mask, orig_latents)
             else:
-                # if no noise provided, noisify unmasked area based on seed(or 0 as fallback)
-                if noise is None:
-                    noise = torch.randn(
-                        orig_latents.shape,
-                        dtype=torch.float32,
-                        device="cpu",
-                        generator=torch.Generator(device="cpu").manual_seed(seed or 0),
-                    ).to(device=orig_latents.device, dtype=orig_latents.dtype)
-
-                    latents = self.scheduler.add_noise(latents, noise, batched_t)
-                    latents = torch.lerp(
-                        orig_latents, latents.to(dtype=orig_latents.dtype), mask.to(dtype=orig_latents.dtype)
-                    )
+                # TODO: debug better with or without Oo
+                if False:
+                    # if no noise provided, noisify unmasked area based on seed(or 0 as fallback)
+                    if noise is None:
+                        noise = torch.randn(
+                            orig_latents.shape,
+                            dtype=torch.float32,
+                            device="cpu",
+                            generator=torch.Generator(device="cpu").manual_seed(seed or 0),
+                        ).to(device=orig_latents.device, dtype=orig_latents.dtype)
+
+                        latents = self.scheduler.add_noise(latents, noise, batched_t)
+                        latents = torch.lerp(
+                            orig_latents, latents.to(dtype=orig_latents.dtype), mask.to(dtype=orig_latents.dtype)
+                        )
 
                 additional_guidance.append(AddsMaskGuidance(mask, orig_latents, self.scheduler, noise))