Merge branch 'main' into xla-autoencoder-qwenimage

Author: sayakpaul

.github/workflows/upload_pr_documentation.yml

@@ -8,7 +8,7 @@ on:
 
 jobs:
   build:
-    uses: huggingface/doc-builder/.github/workflows/upload_pr_documentation.yml@90b4ee2c10b81b5c1a6367c4e6fc9e2fb510a7e3  # main
+    uses: huggingface/doc-builder/.github/workflows/upload_pr_documentation.yml@9ad2de8582b56c017cb530c1165116d40433f1c6  # main
     with:
       package_name: diffusers
     secrets:

src/diffusers/models/transformers/transformer_z_image.py

@@ -777,7 +777,8 @@ def _prepare_sequence(
 
         # Pad token
         feats_cat = torch.cat(feats, dim=0)
-        feats_cat[torch.cat(inner_pad_mask)] = pad_token
+        mask = torch.cat(inner_pad_mask).unsqueeze(-1)
+        feats_cat = torch.where(mask, pad_token, feats_cat)
         feats = list(feats_cat.split(item_seqlens, dim=0))
 
         # RoPE

src/diffusers/pipelines/z_image/pipeline_z_image.py

@@ -486,6 +486,15 @@ def __call__(
         num_warmup_steps = max(len(timesteps) - num_inference_steps * self.scheduler.order, 0)
         self._num_timesteps = len(timesteps)
 
+        # We set the index here to remove DtoH sync, helpful especially during compilation.
+        # Check out more details here: https://github.com/huggingface/diffusers/pull/11696
+        self.scheduler.set_begin_index(0)
+
+        if self.do_classifier_free_guidance and self._cfg_truncation is not None and float(self._cfg_truncation) <= 1:
+            _precomputed_t_norms = ((1000 - timesteps.float()) / 1000).tolist()
+        else:
+            _precomputed_t_norms = None
+
         # 6. Denoising loop
         with self.progress_bar(total=num_inference_steps) as progress_bar:
             for i, t in enumerate(timesteps):
@@ -495,17 +504,9 @@ def __call__(
                 # broadcast to batch dimension in a way that's compatible with ONNX/Core ML
                 timestep = t.expand(latents.shape[0])
                 timestep = (1000 - timestep) / 1000
-                # Normalized time for time-aware config (0 at start, 1 at end)
-                t_norm = timestep[0].item()
-
-                # Handle cfg truncation
                 current_guidance_scale = self.guidance_scale
-                if (
-                    self.do_classifier_free_guidance
-                    and self._cfg_truncation is not None
-                    and float(self._cfg_truncation) <= 1
-                ):
-                    if t_norm > self._cfg_truncation:
+                if _precomputed_t_norms is not None:
+                    if _precomputed_t_norms[i] > self._cfg_truncation:
                         current_guidance_scale = 0.0
 
                 # Run CFG only if configured AND scale is non-zero