few-shots-generation

Few shots image generation based on SinGAN model (See SinGAN Repository: https://github.com/tamarott/SinGAN)

Code

Install dependencies

python -m pip install -r requirements.txt

This code was tested with python 3.7

Train

To train the model on your own batch of images, put the desire training images under a sub-folder of Input/Images/, and run

python main_train.py --input_dir Input/Images/<input_folder>/ --model_name <desired_model_name>

This will also use the resulting trained model to generate random samples starting from the coarsest scale (n=0).

To run this code on a cpu machine, specify --not_cuda when calling main_train.py

Random samples

To generate random samples from any starting generation scale, please first train the model for the desire images (as described above), and then run

python random_samples.py --input_dir Input/Images/<input_folder>/ --model_name <trained_model_name> --mode random_samples --gen_start_scale <generation start scale number>

pay attention: for using the full model, specify the generation start scale to be 0, to start the generation from the second scale, specify it to be 1, and so on.

Quantitative Evaluation

To evaluate the model results with SIFID, run

python /SIFID/sifid_score.py --path2real <path_to_real_images> --path2fake <path_to_fake_images>

For additional informaiton- see SinGAN:

Project | Arxiv | CVF

Official pytorch implementation of the paper: "SinGAN: Learning a Generative Model from a Single Natural Image"

ICCV 2019

SinGAN Repository: https://github.com/tamarott/SinGAN