init commit

.gitignore

@@ -0,0 +1,27 @@
+
+.vscode/
+
+*.pyc
+*pynb*
+
+datasets/
+
+outputs/
+
+samples/
+
+eval/
+
+pretrained_model/
+
+eval/model/
+
+build/
+
+roi_layers.egg-info/
+
+*.so
+
+.ftpconfig
+
+.ftpignore

INSTALL.md

@@ -0,0 +1,26 @@
+## Installation
+
+#### Requiements
+* python3
+* pytorch 1.0
+* numpy
+* matplotlib
+* opencv
+
+```
+pip install -r requiements.txt
+
+# setup for roi_layers
+python setup.py build develop
+```
+
+#### Data Preparation
+Download COCO dataset to `datasets/coco`
+```
+bash scripts/download_coco.sh
+```
+Download VG dataset to `datasets/vg`
+```
+bash scripts/download_vg.sh
+python scripts/preprocess_vg.py
+```

README.md

@@ -0,0 +1,42 @@
+# LostGANs: Image Synthesis From Reconfigurable Layout and Style
+This is implementation of our ICCV19 paper [**Image Synthesis From Reconfigurable Layout and Style**](https://arxiv.org/abs/1908.07500)
+
+## Network Structure
+![network_structure](./figures/network_structure.png)
+
+## Installation
+Check [INSTALL.md](INSTALL.md) for installation instructions.
+#### 1. Download pretrained model
+Download pretrained models to `pretrained_model/`
+* Pretrained model on [COCO](https://drive.google.com/open?id=1WO6fLZqJeTUnmJTmieUopKLj9KGBhGd6)
+* Pretrained model on [VG](https://drive.google.com/open?id=1A_gP_WwZWonlXJhwcdBDgHuVaGSiVjMO)
+
+#### 2. Train models
+```
+python train.py --dataset coco --out_path outputs/
+```
+
+#### 3. Run pretrained model
+```
+python test.py --dataset coco --model_path pretrained_model/G_coco.pth --sample_path samples/coco/
+```
+
+
+## Results
+###### Multiple samples generated from same layout
+![various_out](./figures/various_outs.png)
+###### Generation results by adding new objects or change spatial position of object
+![add_obj](./figures/add_obj.png)
+###### Linear interpolation of instance style
+![style_morph](./figures/style_morph.png)
+###### Synthesized images and learned masks for given layout
+![mask](./figures/mask.png)
+
+## Contact
+Please feel free to report issues and any related problems to Wei Sun (wsun12 at ncsu dot edu) and Tianfu Wu (tianfu_wu at ncsu dot edu).
+
+
+## Reference
+* Synchronized-BatchNorm-PyTorch: [https://github.com/vacancy/Synchronized-BatchNorm-PyTorch](https://github.com/vacancy/Synchronized-BatchNorm-PyTorch)
+* Image Generation from Scene Graphs: [https://github.com/google/sg2im](https://github.com/google/sg2im)
+* Faster R-CNN and Mask R-CNN in PyTorch 1.0: [https://github.com/facebookresearch/maskrcnn-benchmark](https://github.com/facebookresearch/maskrcnn-benchmark)