Closed lixiang007666 closed 2 weeks ago
The changes in this pull request involve modifications to the Attention
class's forward
method in the attention_processor_oflow.py
file. A check has been added to replace the current processor with a default AttnProcessor
if the processor is an instance of DiffusersAttnProcessor
or DiffusersAttnProcessor2_0
. Additionally, necessary import statements for these classes from the diffusers.models.attention_processor
module have been included. The text_to_image_sdxl_enterprise.py
script has also been updated to enhance argument parsing and model compilation options.
File | Change Summary |
---|---|
src/infer_compiler_registry/register_diffusers/attention_processor_oflow.py |
- Modified forward method to include a check for DiffusersAttnProcessor and DiffusersAttnProcessor2_0 .- Added import statements for AttnProcessor and AttnProcessor2_0 . |
onediff_diffusers_extensions/examples/text_to_image_sdxl_enterprise.py |
- Added --compile_text_encoder and --graph arguments in parse_args() function.- Updated logic to handle new arguments for model compilation. |
Objective | Addressed | Explanation |
---|---|---|
Fix TypeError in attention processing (#1105) | ✅ |
🐇 In the meadow, where the bunnies play,
A new processor hops in, to brighten the day.
With attention so keen, it’s ready to flow,
Replacing the old, making magic to show!
Let’s cheer for the change, with a jump and a twirl,
For in this code garden, our dreams can unfurl!
🌼✨
Thank you for using CodeRabbit. We offer it for free to the OSS community and would appreciate your support in helping us grow. If you find it useful, would you consider giving us a shout-out on your favorite social media?
This PR is done:
Summary by CodeRabbit
New Features
diffusers
library.--compile_text_encoder
and--graph
, allowing for more flexible model compilation options.Bug Fixes