Skip to content

fix "Expected all tensors to be on the same device, but found at least two devices" error #11690

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 11 commits into
base: main
Choose a base branch
from
Open

fix "Expected all tensors to be on the same device, but found at least two devices" error #11690

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
11 commits
Select commit Hold shift + click to select a range
6001899
xx
yao-matrix Jun 10, 2025
8a1d6e5
fix
yao-matrix Jun 11, 2025
603257b
Update model_loading_utils.py
yao-matrix Jun 11, 2025
8cdfdd8
Update test_models_unet_2d_condition.py
yao-matrix Jun 11, 2025
45e29bd
Update test_models_unet_2d_condition.py
yao-matrix Jun 11, 2025
2a7c17d
Merge branch 'main' into xpu
yao-matrix Jun 11, 2025
fae7c70
fix style
yao-matrix Jun 11, 2025
80fdbfc
Merge branch 'main' into xpu
yao-matrix Jun 11, 2025
97a37a1
Merge branch 'main' into xpu
yao-matrix Jun 11, 2025
5f0c794
Merge branch 'main' into xpu
yao-matrix Jun 12, 2025
8cd06b3
Merge branch 'main' into xpu
yao-matrix Jun 13, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion src/diffusers/models/unets/unet_2d_blocks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2557,7 +2557,8 @@ def forward(
b1=self.b1,
b2=self.b2,
)

if hidden_states.device != res_hidden_states.device:
res_hidden_states = res_hidden_states.to(hidden_states.device)
Comment on lines +2560 to +2561
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we shouldn't need that since both hidden_states and res_hidden_states should be on the same device no ? The pre-forward hook added by accelerate should be move all the inputs to the same device.

hidden_states = torch.cat([hidden_states, res_hidden_states], dim=1)

if torch.is_grad_enabled() and self.gradient_checkpointing:
Expand Down
5 changes: 2 additions & 3 deletions tests/models/test_modeling_common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -70,7 +70,6 @@
require_torch_2,
require_torch_accelerator,
require_torch_accelerator_with_training,
require_torch_gpu,
require_torch_multi_accelerator,
run_test_in_subprocess,
slow,
Expand Down Expand Up @@ -1744,7 +1743,7 @@ def test_push_to_hub_library_name(self):
delete_repo(self.repo_id, token=TOKEN)


@require_torch_gpu
@require_torch_accelerator
@require_torch_2
@is_torch_compile
@slow
Expand Down Expand Up @@ -1789,7 +1788,7 @@ def test_compile_with_group_offloading(self):
model.eval()
# TODO: Can test for other group offloading kwargs later if needed.
group_offload_kwargs = {
"onload_device": "cuda",
"onload_device": torch_device,
"offload_device": "cpu",
"offload_type": "block_level",
"num_blocks_per_group": 1,
Expand Down