[Enhance] Improve powerpaint (#2080)

* add readme

* add readme

* add require

* fix readme

* fix according to readme

* fix lint

projects/powerpaint/README.md

@@ -0,0 +1,60 @@
+# A Task is Worth One Word: Learning with Task Prompts for High-Quality Versatile Image Inpainting
+
+### [Project Page](https://powerpaint.github.io/) | [Paper](https://arxiv.org/abs/2312.03594)
+
+This README provides a step-by-step guide to download the repository, set up the required virtual environment named "PowerPaint" using conda, and run PowerPaint with or without ControlNet.
+
+**Stronger Model Weights and Online Demo Coming Soon！**
+
+<img src='https://github.com/open-mmlab/mmagic/assets/12782558/acd01391-c73f-4997-aafd-0869aebcc915'/>
+
+## Getting Started
+
+```bash
+# Clone the Repository
+git clone https://github.com/open-mmlab/mmagic.git
+
+# Navigate to the Repository
+cd projects/powerpaint
+
+# Create Virtual Environment with Conda
+conda create --name PowerPaint python=3.9
+conda activate PowerPaint
+
+# Install Dependencies
+pip install -r requirements.txt
+
+# Create Models Folder
+mkdir models
+
+# Set up Git LFS
+git lfs install
+
+# Clone PowerPaint Model
+git lfs clone https://huggingface.co/JunhaoZhuang/PowerPaint-v1/ ./models
+```
+
+## Run PowerPaint
+
+To run PowerPaint, execute the following command:
+
+```bash
+python gradio_PowerPaint.py
+```
+
+This command will launch the Gradio interface for PowerPaint.
+
+Feel free to explore and create stunning images with PowerPaint!
+
+## BibTeX
+
+```
+@misc{zhuang2023task,
+      title={A Task is Worth One Word: Learning with Task Prompts for High-Quality Versatile Image Inpainting},
+      author={Junhao Zhuang and Yanhong Zeng and Wenran Liu and Chun Yuan and Kai Chen},
+      year={2023},
+      eprint={2312.03594},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV}
+}
+```

projects/powerpaint/gradio_PowerPaint.py

@@ -34,9 +34,9 @@
     initialize_tokens=['a', 'a', 'a'],
     num_vectors_per_token=10)
 pipe.unet.load_state_dict(
-    torch.load('./models/unet/diffusion_pytorch_model.bin'), strict=False)
+    torch.load('./models/diffusion_pytorch_model.bin'), strict=False)
 pipe.text_encoder.load_state_dict(
-    torch.load('./models/text_encoder/pytorch_model.bin'), strict=False)
+    torch.load('./models/pytorch_model.bin'), strict=False)
 pipe = pipe.to('cuda')
 
 depth_estimator = DPTForDepthEstimation.from_pretrained(

projects/powerpaint/requirements.txt

@@ -0,0 +1,8 @@
+controlnet-aux==0.0.3
+diffusers==0.23.1
+gradio==3.23.0
+mmengine
+opencv-python
+torch
+torchvision
+transformers

projects/powerpaint/utils/utils.py

@@ -1,4 +1,5 @@
 import copy
+import importlib
 import os
 import random
 from logging import WARNING
@@ -8,7 +9,21 @@
 import torch.nn as nn
 from mmengine import print_log
 
-from mmagic.utils import try_import
+
+def try_import(name: str):
+    """Try to import a module.
+
+    Args:
+        name (str): Specifies what module to import in absolute or relative
+            terms (e.g. either pkg.mod or ..mod).
+    Returns:
+        ModuleType or None: If importing successfully, returns the imported
+        module, otherwise returns None.
+    """
+    try:
+        return importlib.import_module(name)
+    except ImportError:
+        return None
 
 
 class TokenizerWrapper: