Skip to content

Improve pos embed for Flux.1 inference on Ascend NPU #12534

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 1 commit into
base: main
Choose a base branch
from
+6 −2
Open

Improve pos embed for Flux.1 inference on Ascend NPU #12534

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
8 changes: 6 additions & 2 deletions src/diffusers/models/transformers/transformer_flux.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,7 +22,7 @@

from ...configuration_utils import ConfigMixin, register_to_config
from ...loaders import FluxTransformer2DLoadersMixin, FromOriginalModelMixin, PeftAdapterMixin
from ...utils import USE_PEFT_BACKEND, logging, scale_lora_layers, unscale_lora_layers
from ...utils import USE_PEFT_BACKEND, is_torch_npu_available, logging, scale_lora_layers, unscale_lora_layers
from ...utils.torch_utils import maybe_allow_in_graph
from .._modeling_parallel import ContextParallelInput, ContextParallelOutput
from ..attention import AttentionMixin, AttentionModuleMixin, FeedForward
Expand Down Expand Up @@ -717,7 +717,11 @@ def forward(
img_ids = img_ids[0]

ids = torch.cat((txt_ids, img_ids), dim=0)
image_rotary_emb = self.pos_embed(ids)
if is_torch_npu_available():
freqs_cos, freqs_sin = self.pos_embed(ids.cpu())
image_rotary_emb = (freqs_cos.npu(), freqs_sin.npu())
else:
image_rotary_emb = self.pos_embed(ids)

if joint_attention_kwargs is not None and "ip_adapter_image_embeds" in joint_attention_kwargs:
ip_adapter_image_embeds = joint_attention_kwargs.pop("ip_adapter_image_embeds")
Expand Down