zideliu / StyleDrop-PyTorch

Unoffical implement for [StyleDrop](https://arxiv.org/abs/2306.00983)
MIT License
552 stars 26 forks source link

python extract_empty_feature.py #17

Open wpl427 opened 11 months ago

wpl427 commented 11 months ago

. ├── assets │ ├── ckpts │ │ ├── cc3m-285000.ckpt │ │ │ ├── lr_scheduler.pth │ │ │ ├── nnet_ema.pth │ │ │ ├── nnet.pth │ │ │ ├── optimizer.pth │ │ │ └── step.pth │ │ └── imagenet256-450000.ckpt │ │ ├── lr_scheduler.pth │ │ ├── nnet_ema.pth │ │ ├── nnet.pth │ │ ├── optimizer.pth │ │ └── step.pth │ ├── fid_stats │ │ ├── fid_stats_cc3m_val.npz │ │ └── fid_stats_imagenet256_guided_diffusion.npz │ ├── pipeline.png | ├── contexts │ │ └── empty_context.npy └── └── vqgan_jax_strongaug.ckpt python extract_empty_feature.py The fid stats folder and the following two files are missing

. ├── assets │ ├── ckpts │ │ ├── cc3m-285000.ckpt │ │ │ ├── lr_scheduler.pth │ │ │ ├── nnet_ema.pth │ │ │ ├── nnet.pth │ │ │ ├── optimizer.pth │ │ │ └── step.pth │ │ └── imagenet256-450000.ckpt │ │ ├── lr_scheduler.pth │ │ ├── nnet_ema.pth │ │ ├── nnet.pth │ │ ├── optimizer.pth │ │ └── step.pth │ ├── pipeline.png | ├── contexts │ │ └── empty_context.npy └── └── vqgan_jax_strongaug.ckpt

zideliu commented 10 months ago

Download the reference statistics for FID from this link. Place the downloaded .npz file in assets/fid_stats.

But actually, they are not used