Home

Awesome

<p align=center>PixelFace</p>

<!-- [![Paper](http://img.shields.io/badge/paper-arxiv.2010.04513-blue.svg)]() -->

License: MIT

Dataset

MMCelebA-HQ: image, pickle, json

Pretraiend Models

  1. Text Encoder
  2. Image Encoder
  3. CheckPoint

Train on Multi-Modal CelebA-HQ

CUDA_VISIBLE_DEVICES=0,1 python -m torch.distributed.launch --nproc_per_node NPROC_PER_NODE --master_port MASTER_PORT main.py --cfg cfg/mmceleba.yml

or

CUDA_VISIBLE_DEVICES=0,1 torchrun --nproc_per_node NPROC_PER_NODE --master_port MASTER_PORT main.py --cfg cfg/mmceleba.yml

Test

CUDA_VISIBLE_DEVICES=0,1 python -m torch.distributed.launch --nproc_per_node NPROC_PER_NODE --master_port MASTER_PORT main.py --cfg cfg/eval_mmceleba.yml

or

CUDA_VISIBLE_DEVICES=0,1 torchrun --nproc_per_node NPROC_PER_NODE --master_port MASTER_PORT main.py --cfg cfg/eval_mmceleba.yml