update

Signed-off-by: Liu, Kaixuan <kaixuan.liu@intel.com>

examples/flux-control/train_control_flux.py

@@ -1105,7 +1105,7 @@ def main(args):
 
                 # text encoding.
                 captions = batch["captions"]
-                text_encoding_pipeline = text_encoding_pipeline.to("cuda")
+                text_encoding_pipeline = text_encoding_pipeline.to(accelerator.device)
                 with torch.no_grad():
                     prompt_embeds, pooled_prompt_embeds, text_ids = text_encoding_pipeline.encode_prompt(
                         captions, prompt_2=None

examples/flux-control/train_control_lora_flux.py

@@ -1251,7 +1251,7 @@ def main(args):
 
                 # text encoding.
                 captions = batch["captions"]
-                text_encoding_pipeline = text_encoding_pipeline.to("cuda")
+                text_encoding_pipeline = text_encoding_pipeline.to(accelerator.device)
                 with torch.no_grad():
                     prompt_embeds, pooled_prompt_embeds, text_ids = text_encoding_pipeline.encode_prompt(
                         captions, prompt_2=None

src/diffusers/hooks/group_offloading.py

@@ -22,7 +22,7 @@ from typing import Set
 import safetensors.torch
 import torch
 
-from ..utils import get_logger, is_accelerate_available, is_torchao_available
+from ..utils import get_logger, is_accelerate_available
 from ._common import _GO_LC_SUPPORTED_PYTORCH_LAYERS
 from .hooks import HookRegistry, ModelHook
 
@@ -35,54 +35,6 @@ if is_accelerate_available():
 logger = get_logger(__name__)  # pylint: disable=invalid-name
 
 
-def _is_torchao_tensor(tensor: torch.Tensor) -> bool:
-    if not is_torchao_available():
-        return False
-    from torchao.utils import TorchAOBaseTensor
-
-    return isinstance(tensor, TorchAOBaseTensor)
-
-
-def _get_torchao_inner_tensor_names(tensor: torch.Tensor) -> list[str]:
-    """Get names of all internal tensor data attributes from a TorchAO tensor."""
-    cls = type(tensor)
-    names = list(getattr(cls, "tensor_data_names", []))
-    for attr_name in getattr(cls, "optional_tensor_data_names", []):
-        if getattr(tensor, attr_name, None) is not None:
-            names.append(attr_name)
-    return names
-
-
-def _swap_torchao_tensor(param: torch.Tensor, source: torch.Tensor) -> None:
-    """Move a TorchAO parameter to the device of `source` via `swap_tensors`.
-
-    `param.data = source` does not work for `_make_wrapper_subclass` tensors because the `.data` setter only replaces
-    the outer wrapper storage while leaving the subclass's internal attributes (e.g. `.qdata`, `.scale`) on the
-    original device. `swap_tensors` swaps the full tensor contents in-place, preserving the parameter's identity so
-    that any dict keyed by `id(param)` remains valid.
-
-    Refer to https://github.com/huggingface/diffusers/pull/13276#discussion_r2944471548 for the full discussion.
-    """
-    torch.utils.swap_tensors(param, source)
-
-
-def _restore_torchao_tensor(param: torch.Tensor, source: torch.Tensor) -> None:
-    """Restore internal tensor data of a TorchAO parameter from `source` without mutating `source`.
-
-    Unlike `_swap_torchao_tensor` this copies attribute references one-by-one via `setattr` so that `source` is **not**
-    modified. Use this when `source` is a cached tensor that must remain unchanged (e.g. a pinned CPU copy in
-    `cpu_param_dict`).
-    """
-    for attr_name in _get_torchao_inner_tensor_names(source):
-        setattr(param, attr_name, getattr(source, attr_name))
-
-
-def _record_stream_torchao_tensor(param: torch.Tensor, stream) -> None:
-    """Record stream for all internal tensors of a TorchAO parameter."""
-    for attr_name in _get_torchao_inner_tensor_names(param):
-        getattr(param, attr_name).record_stream(stream)
-
-
 # fmt: off
 _GROUP_OFFLOADING = "group_offloading"
 _LAYER_EXECUTION_TRACKER = "layer_execution_tracker"
@@ -205,16 +157,9 @@ class ModuleGroup:
             pinned_dict = None
 
     def _transfer_tensor_to_device(self, tensor, source_tensor, default_stream):
-        moved = source_tensor.to(self.onload_device, non_blocking=self.non_blocking)
-        if _is_torchao_tensor(tensor):
-            _swap_torchao_tensor(tensor, moved)
-        else:
-            tensor.data = moved
+        tensor.data = source_tensor.to(self.onload_device, non_blocking=self.non_blocking)
         if self.record_stream:
-            if _is_torchao_tensor(tensor):
-                _record_stream_torchao_tensor(tensor, default_stream)
-            else:
-                tensor.data.record_stream(default_stream)
+            tensor.data.record_stream(default_stream)
 
     def _process_tensors_from_modules(self, pinned_memory=None, default_stream=None):
         for group_module in self.modules:
@@ -300,35 +245,18 @@ class ModuleGroup:
 
             for group_module in self.modules:
                 for param in group_module.parameters():
-                    if _is_torchao_tensor(param):
-                        _restore_torchao_tensor(param, self.cpu_param_dict[param])
-                    else:
-                        param.data = self.cpu_param_dict[param]
-            for param in self.parameters:
-                if _is_torchao_tensor(param):
-                    _restore_torchao_tensor(param, self.cpu_param_dict[param])
-                else:
                     param.data = self.cpu_param_dict[param]
+            for param in self.parameters:
+                param.data = self.cpu_param_dict[param]
             for buffer in self.buffers:
-                if _is_torchao_tensor(buffer):
-                    _restore_torchao_tensor(buffer, self.cpu_param_dict[buffer])
-                else:
-                    buffer.data = self.cpu_param_dict[buffer]
+                buffer.data = self.cpu_param_dict[buffer]
         else:
             for group_module in self.modules:
                 group_module.to(self.offload_device, non_blocking=False)
             for param in self.parameters:
-                if _is_torchao_tensor(param):
-                    moved = param.data.to(self.offload_device, non_blocking=False)
-                    _swap_torchao_tensor(param, moved)
-                else:
-                    param.data = param.data.to(self.offload_device, non_blocking=False)
+                param.data = param.data.to(self.offload_device, non_blocking=False)
             for buffer in self.buffers:
-                if _is_torchao_tensor(buffer):
-                    moved = buffer.data.to(self.offload_device, non_blocking=False)
-                    _swap_torchao_tensor(buffer, moved)
-                else:
-                    buffer.data = buffer.data.to(self.offload_device, non_blocking=False)
+                buffer.data = buffer.data.to(self.offload_device, non_blocking=False)
 
     @torch.compiler.disable()
     def onload_(self):

tests/models/transformers/test_models_transformer_glm_image.py

@@ -0,0 +1,99 @@
+# coding=utf-8
+# Copyright 2025 HuggingFace Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import torch
+
+from diffusers import GlmImageTransformer2DModel
+from diffusers.utils.torch_utils import randn_tensor
+
+from ...testing_utils import enable_full_determinism, torch_device
+from ..testing_utils import (
+    BaseModelTesterConfig,
+    ModelTesterMixin,
+    TorchCompileTesterMixin,
+    TrainingTesterMixin,
+)
+
+
+enable_full_determinism()
+
+
+class GlmImageTransformerTesterConfig(BaseModelTesterConfig):
+    @property
+    def model_class(self):
+        return GlmImageTransformer2DModel
+
+    @property
+    def main_input_name(self) -> str:
+        return "hidden_states"
+
+    @property
+    def output_shape(self) -> tuple:
+        return (16, 8, 8)
+
+    @property
+    def input_shape(self) -> tuple:
+        return (4, 8, 8)
+
+    @property
+    def generator(self):
+        return torch.Generator("cpu").manual_seed(0)
+
+    def get_init_dict(self) -> dict:
+        return {
+            "patch_size": 2,
+            "in_channels": 4,
+            "out_channels": 4,
+            "num_layers": 1,
+            "attention_head_dim": 8,
+            "num_attention_heads": 2,
+            "text_embed_dim": 32,
+            "time_embed_dim": 16,
+            "condition_dim": 8,
+            "prior_vq_quantizer_codebook_size": 64,
+        }
+
+    def get_dummy_inputs(self, batch_size: int = 1) -> dict[str, torch.Tensor]:
+        num_channels = 4
+        height = width = 8
+        sequence_length = 12
+
+        return {
+            "hidden_states": randn_tensor(
+                (batch_size, num_channels, height, width), generator=self.generator, device=torch_device
+            ),
+            "encoder_hidden_states": randn_tensor(
+                (batch_size, sequence_length, 32), generator=self.generator, device=torch_device
+            ),
+            "prior_token_id": torch.randint(0, 64, size=(batch_size,), generator=self.generator).to(torch_device),
+            "prior_token_drop": torch.zeros(batch_size, dtype=torch.bool, device=torch_device),
+            "timestep": torch.randint(0, 1000, size=(batch_size,), generator=self.generator).to(torch_device),
+            "target_size": torch.tensor([[height, width]] * batch_size, dtype=torch.float32).to(torch_device),
+            "crop_coords": torch.tensor([[0, 0]] * batch_size, dtype=torch.float32).to(torch_device),
+        }
+
+
+class TestGlmImageTransformer(GlmImageTransformerTesterConfig, ModelTesterMixin):
+    pass
+
+
+class TestGlmImageTransformerTraining(GlmImageTransformerTesterConfig, TrainingTesterMixin):
+    def test_gradient_checkpointing_is_applied(self):
+        expected_set = {"GlmImageTransformer2DModel"}
+        super().test_gradient_checkpointing_is_applied(expected_set=expected_set)
+
+
+class TestGlmImageTransformerCompile(GlmImageTransformerTesterConfig, TorchCompileTesterMixin):
+    pass