/Text2Image-Synthesis

GAN based Text2Image-Synthesis

Primary LanguagePython

Text2Image-Synthesis

C-GAN, TextEncoder, COCO-2014

Install axel download accelerator

$ sudo apt-get install axel
$ axel -n 10 http://images.cocodataset.org/zips/train2014.zip
$ axel -n 10 http://images.cocodataset.org/zips/val2014.zip
$ axel -n 10 http://images.cocodataset.org/zips/test2014.zip
$ axel -n 10 http://images.cocodataset.org/annotations/annotations_trainval2014.zip
$ axel -n 10 http://images.cocodataset.org/annotations/image_info_test2014.zip

Create folders

$ mkdir coco-2014
$ mkdir coco-2014/images/
$ mkdir coco-2014/resized/
$ mkdir coco-2014/annotations/

Run python script

$ pip3 install -r requirements.txt
$ python3 utils/resize_image.py --source=coco-2014/images --target=coco-2014/resized/
$ python3 run_with_gpu.py --root=../coco-2014 --sample_interval=20 --cuda_id=1 --batch_size=128