This repository provides the official Pytorch implementation of our paper: FEW-SHOT KNOWLEDGE TRANSFER FOR FINE-GRAINED CARTOON FACE GENERATION (ICME 21 oral ).
We released the pre-trained single model and multi-group model so that people could test more widely. Please download the pre-trained models from [Google Driver]
Please download the dataset from [Google Driver] and extract it to
dataset
folder.
We train our model on a server with a NVIDIA Tesla V100 of 32G memory.
python train.py
If the memory of gpu is not sufficient, set --light
to True
--light
to Falsepython test.py --dataset {folder_for_dataset} --model_path {path_to_model_ckpt} --remark {exp_name} --label {0, 1, 2, 3 for specific groups}
Our codes refer to Project Photo2cartoon