Skip to content

device_map in load_model_dict_into_meta #10851

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

Merged
merged 4 commits into from
Feb 21, 2025
Merged

device_map in load_model_dict_into_meta #10851

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
18454cf
`device_map` in `load_model_dict_into_meta`
hlky Feb 21, 2025
040d657
_LOW_CPU_MEM_USAGE_DEFAULT
hlky Feb 21, 2025
9e2e165
Merge branch 'main' into model-loading-refactor-fixes
hlky Feb 21, 2025
7d96d88
fix is_peft_version is_bitsandbytes_version
hlky Feb 21, 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
7 changes: 4 additions & 3 deletions src/diffusers/loaders/transformer_flux.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -82,7 +82,8 @@ def _convert_ip_adapter_image_proj_to_diffusers(self, state_dict, low_cpu_mem_us
if not low_cpu_mem_usage:
image_projection.load_state_dict(updated_state_dict, strict=True)
else:
load_model_dict_into_meta(image_projection, updated_state_dict, device=self.device, dtype=self.dtype)
device_map = {"": self.device}
load_model_dict_into_meta(image_projection, updated_state_dict, device_map=device_map, dtype=self.dtype)

return image_projection

Expand Down Expand Up @@ -151,9 +152,9 @@ def _convert_ip_adapter_attn_to_diffusers(self, state_dicts, low_cpu_mem_usage=F
if not low_cpu_mem_usage:
attn_procs[name].load_state_dict(value_dict)
else:
device = self.device
device_map = {"": self.device}
dtype = self.dtype
load_model_dict_into_meta(attn_procs[name], value_dict, device=device, dtype=dtype)
load_model_dict_into_meta(attn_procs[name], value_dict, device_map=device_map, dtype=dtype)

key_id += 1

Expand Down
6 changes: 4 additions & 2 deletions src/diffusers/loaders/transformer_sd3.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,8 +75,9 @@ def _convert_ip_adapter_attn_to_diffusers(
if not low_cpu_mem_usage:
attn_procs[name].load_state_dict(layer_state_dict[idx], strict=True)
else:
device_map = {"": self.device}
load_model_dict_into_meta(
attn_procs[name], layer_state_dict[idx], device=self.device, dtype=self.dtype
attn_procs[name], layer_state_dict[idx], device_map=device_map, dtype=self.dtype
)

return attn_procs
Expand Down Expand Up @@ -144,7 +145,8 @@ def _convert_ip_adapter_image_proj_to_diffusers(
if not low_cpu_mem_usage:
image_proj.load_state_dict(updated_state_dict, strict=True)
else:
load_model_dict_into_meta(image_proj, updated_state_dict, device=self.device, dtype=self.dtype)
device_map = {"": self.device}
load_model_dict_into_meta(image_proj, updated_state_dict, device_map=device_map, dtype=self.dtype)

return image_proj

Expand Down
6 changes: 4 additions & 2 deletions src/diffusers/loaders/unet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -753,7 +753,8 @@ def _convert_ip_adapter_image_proj_to_diffusers(self, state_dict, low_cpu_mem_us
if not low_cpu_mem_usage:
image_projection.load_state_dict(updated_state_dict, strict=True)
else:
load_model_dict_into_meta(image_projection, updated_state_dict, device=self.device, dtype=self.dtype)
device_map = {"": self.device}
load_model_dict_into_meta(image_projection, updated_state_dict, device_map=device_map, dtype=self.dtype)

return image_projection

Expand Down Expand Up @@ -846,7 +847,8 @@ def _convert_ip_adapter_attn_to_diffusers(self, state_dicts, low_cpu_mem_usage=F
else:
device = next(iter(value_dict.values())).device
dtype = next(iter(value_dict.values())).dtype
load_model_dict_into_meta(attn_procs[name], value_dict, device=device, dtype=dtype)
device_map = {"": device}
load_model_dict_into_meta(attn_procs[name], value_dict, device_map=device_map, dtype=dtype)

key_id += 2

Expand Down