update readme

README.md

@@ -17,11 +17,12 @@ license: mit
 Star ⭐ us if you like it!
 
 ## News
-- 02/Jan/2025, Update the mask generator to improve results. Add ref unet acceleration, boosting prediction speed by 30%. Include more controls in Advanced Options to enhance user experience. Enable intermediate result output for easier development. Enjoy using it!
-- 18/Dec/2024, thanks to @[StartHua](https://github.com/StartHua) for integrating Leffa into ComfyUI! Here is the [repo](https://github.com/StartHua/Comfyui_leffa)!
-- 16/Dec/2024, the virtual try-on [model](https://huggingface.co/franciszzj/Leffa/blob/main/virtual_tryon_dc.pth) trained on DressCode is released.
-- 12/Dec/2024, the HuggingFace [demo](https://huggingface.co/spaces/franciszzj/Leffa) and [models](https://huggingface.co/franciszzj/Leffa) (virtual try-on model trained on VITON-HD and pose transfer model trained on DeepFashion) are released.
-- 11/Dec/2024, the [arXiv](https://arxiv.org/abs/2412.08486) version of the paper is released.
+- 09/Jan/2025. Inference defaults to float16, generating an image in 6 seconds (on A100).
+- 02/Jan/2025. Update the mask generator to improve results. Add ref unet acceleration, boosting prediction speed by 30%. Include more controls in Advanced Options to enhance user experience. Enable intermediate result output for easier development. Enjoy using it!
+- 18/Dec/2024. Thanks to @[StartHua](https://github.com/StartHua) for integrating Leffa into ComfyUI! Here is the [repo](https://github.com/StartHua/Comfyui_leffa)!
+- 16/Dec/2024. The virtual try-on [model](https://huggingface.co/franciszzj/Leffa/blob/main/virtual_tryon_dc.pth) trained on DressCode is released.
+- 12/Dec/2024. The HuggingFace [demo](https://huggingface.co/spaces/franciszzj/Leffa) and [models](https://huggingface.co/franciszzj/Leffa) (virtual try-on model trained on VITON-HD and pose transfer model trained on DeepFashion) are released.
+- 11/Dec/2024. The [arXiv](https://arxiv.org/abs/2412.08486) version of the paper is released.
 
 
 *[Leffa](https://en.wiktionary.org/wiki/leffa)* is a unified framework for controllable person image generation that enables precise manipulation of both appearance (i.e., virtual try-on) and pose (i.e., pose transfer).

app.py

@@ -163,9 +163,13 @@ if __name__ == "__main__":
     garment_images = list_dir(f"{example_dir}/garment")
 
     title = "## Leffa: Learning Flow Fields in Attention for Controllable Person Image Generation"
-    link = "[📚 Paper](https://arxiv.org/abs/2412.08486) - [🤖 Code](https://github.com/franciszzj/Leffa) - [🔥 Demo](https://huggingface.co/spaces/franciszzj/Leffa) - [🤗 Model](https://huggingface.co/franciszzj/Leffa)"
+    link = """[📚 Paper](https://arxiv.org/abs/2412.08486) - [🤖 Code](https://github.com/franciszzj/Leffa) - [🔥 Demo](https://huggingface.co/spaces/franciszzj/Leffa) - [🤗 Model](https://huggingface.co/franciszzj/Leffa)
+           
+           Star ⭐ us if you like it!
+           """
     news = """## News
-            - 02/Jan/2025, Update the mask generator to improve results. Add ref unet acceleration, boosting prediction speed by 30%. Include more controls in Advanced Options to enhance user experience. Enable intermediate result output for easier development. Enjoy using it!
+            - 09/Jan/2025. Inference defaults to float16, generating an image in 6 seconds (on A100).
+
             More news can be found in the [GitHub repository](https://github.com/franciszzj/Leffa).
             """
     description = "Leffa is a unified framework for controllable person image generation that enables precise manipulation of both appearance (i.e., virtual try-on) and pose (i.e., pose transfer)."