fix(optimize): use fp16 as default

.gitignore

@@ -170,4 +170,5 @@ lightning_logs/*
 /tmp_select/
 /tmp_select/*
 /logs/
-/logs/*
+/logs/*
+_Record.md

Examples.md

@@ -161,9 +161,3 @@ expressive eyes. <br/>
 ````shell
 python svgdreamer.py x=painting "prompt='self portrait of Van Gogh. oil painting. cmyk portrait. multi colored. defiant and beautiful. cmyk. expressive eyes.'" x.num_paths=256 result_path='./logs/VanGogh-Portrait'
 ````
-
-### Case: planet Saturn
-
-```shell
-python svgdreamer.py x=iconography-s1 skip_sive=False "prompt='An icon of the planet Saturn. minimal flat 2D vector icon. plain color background. trending on ArtStation.'" token_ind=6 x.sive.bg.num_iter=50 x.sive.fg.num_iter=50 x.vpsd.t_schedule='randint' result_path='./logs/Saturn' multirun=True state.mprec='fp16
-```

README.md

@@ -1,7 +1,11 @@
 # SVGDreamer: Text Guided SVG Generation with Diffusion Model
 
+<div align="center" style="line-height: 1;">
+
 [![CVPR 2024](https://img.shields.io/badge/CVPR%202024-Paper-4169E1?style=for-the-badge&logo=arxiv&logoColor=white)](https://arxiv.org/abs/2312.16476) [![arXiv](https://img.shields.io/badge/arXiv-2312.16476-8A2BE2?style=for-the-badge&logo=arxiv&logoColor=white)](https://arxiv.org/abs/2312.16476) [![Project Website](https://img.shields.io/badge/Website-Project%20Page-4682B4?style=for-the-badge&logo=github&logoColor=white)](https://ximinng.github.io/SVGDreamer-project/) [![English Blog](https://img.shields.io/badge/Blog-English-00CED1?style=for-the-badge&logo=huggingface&logoColor=white)](https://huggingface.co/blog/xingxm/svgdreamer) [![中文博客](https://img.shields.io/badge/博客-中文-1E90FF?style=for-the-badge&logo=zhihu&logoColor=white)](https://zhuanlan.zhihu.com/p/687525994)
 
+</div>
+
 This repository contains the official implementation of our CVPR 2024 paper, "SVGDreamer: Text-Guided SVG Generation
 with Diffusion Model." The method leverages a diffusion-based approach to produce high-quality SVGs guided by text
 prompts.
@@ -60,7 +64,7 @@ If you prefer manual setup, download the model from Hugging Face:
 
 The model will be stored at:
 
-📁 Default Path: `/home/user/.cache/huggingface/hub/models--stabilityai--stable-diffusion-2-1-base`
+📁 Default Path: `~/.cache/huggingface/hub/models--stabilityai--stable-diffusion-2-1-base`
 
 ## 🔥 Quickstart: synthesize **6** SVGs at once
 
@@ -255,7 +259,7 @@ We gratefully thank the authors for their wonderful works.
 If you use this code for your research, please cite the following work:
 
 ```
-@InProceedings{svgdreamer_xing_2023,
+@InProceedings{xing2024svgdreamer,
     author    = {Xing, Ximing and Zhou, Haitao and Wang, Chuang and Zhang, Jing and Xu, Dong and Yu, Qian},
     title     = {SVGDreamer: Text Guided SVG Generation with Diffusion Model},
     booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},

conf/config.yaml

@@ -11,7 +11,7 @@ skip_sive: True # optimize from scratch without SIVE init
 # Accelerate config
 state:
   cpu: False # use cpu
-  mprec: 'no' # mixed precision, choices: 'no', 'fp16', 'bf16'
+  mprec: 'fp16' # mixed precision, choices: 'no', 'fp16', 'bf16'
 
 # Diffusers config
 diffuser:

svgdreamer/painter/VPSD_pipeline.py

@@ -11,8 +11,8 @@ import torch.nn.functional as F
 from torchvision import transforms
 from diffusers import StableDiffusionPipeline, UNet2DConditionModel
 from diffusers import DDIMScheduler
-from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion import (
-    rescale_noise_cfg, StableDiffusionPipelineOutput)
+from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion import rescale_noise_cfg, \
+    StableDiffusionPipelineOutput
 from diffusers.models.attention_processor import LoRAAttnProcessor
 from diffusers.loaders import AttnProcsLayers
 

svgdreamer/painter/painter_params.py

@@ -320,7 +320,7 @@ class Painter(DiffVGState):
                 select_path_ids_by_opc = get_keys_below_threshold(opacity_record_, opacity_threshold)
 
                 if len(select_path_ids_by_opc) > 0:
-                    print("-> opacity_record: ", [f"{k}: {v:.3f}" for k, v in opacity_record_.items()])
+                    # print("-> opacity_record: ", [f"{k}: {v:.3f}" for k, v in opacity_record_.items()])
                     print("select_path_ids_by_opc: ", select_path_ids_by_opc)
                 else:
                     stats_np = np.array(list(opacity_record_.values()))
@@ -335,7 +335,7 @@ class Painter(DiffVGState):
                         select_path_ids_by_area.append(shape.id)
 
                 if len(select_path_ids_by_area) > 0:
-                    print("-> area_records: ", ['%.2f' % i for i in area_records])
+                    # print("-> area_records: ", ['%.2f' % i for i in area_records])
                     print("select_path_ids_by_area: ", select_path_ids_by_area)
                 else:
                     stats_np = np.array(area_records)
@@ -349,7 +349,7 @@ class Painter(DiffVGState):
                 select_path_ids_by_opc = get_keys_below_threshold(opacity_record_, opacity_threshold)
 
                 if len(select_path_ids_by_opc) > 0:
-                    print("-> opacity_record: ", [f"{k}: {v:.3f}" for k, v in opacity_record_.items()])
+                    # print("-> opacity_record: ", [f"{k}: {v:.3f}" for k, v in opacity_record_.items()])
                     print("select_path_ids_by_opc: ", select_path_ids_by_opc)
                 else:
                     stats_np = np.array(list(opacity_record_.values()))