forked from leejet/stable-diffusion.cpp
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
彭旭蕊
committed
Oct 30, 2024
1 parent
ac54e00
commit 61ab316
Showing
1 changed file
with
35 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,35 @@ | ||
| ## Using TAESD & TAESDXL to faster decoding | ||
|
|
||
| We'd like to introduce our recent work [Hybrid-SD](https://github.com/bytedance/Hybrid-SD.git). It provides a [Hybrid-sd-tinyvae](https://huggingface.co/cqyan/hybrid-sd-tinyvae) TinyAutoencoder and a [Hybrid-sd-tinyvae-xl](https://huggingface.co/cqyan/hybrid-sd-tinyvae-xl) TinyAutoencoder, which are finetuned version of [TAESD](https://huggingface.co/madebyollin/taesd) and [TAESDXL](https://huggingface.co/madebyollin/taesdxl) respectively. We aim to fix the low-saturation and low-resolution problem in SD1.x and SDXL, while narrowing the loss from using smaller decoders. You can use our models to accelerate the decoding of latent images in base model SD1.x or SDXL by following these steps: | ||
|
|
||
|
|
||
| ### Hybrid-sd-tinyvae for SD1.x | ||
| Download the model weight [Hybrid-sd-tinyvae](https://huggingface.co/cqyan/hybrid-sd-tinyvae) | ||
| ```bash | ||
| curl -L -O https://huggingface.co/cqyan/hybrid-sd-tinyvae/blob/main/diffusion_pytorch_model.safetensors | ||
| ``` | ||
| T2I Comparison using one A100 GPU, The image order from left to right : SD1.4-VAE -> TAESD -> Hybrid-sd-tinyvae | ||
|  | ||
|  | ||
| Specify the model path using the `--taesd PATH` parameter. example: | ||
| ```bash | ||
| sd -m ../models/v1-5-pruned-emaonly.safetensors -p "a lovely cat" --taesd ../models/diffusion_pytorch_model.safetensors | ||
| ``` | ||
| <br /> | ||
| <br /> | ||
|
|
||
|
|
||
|
|
||
|
|
||
| ### Hybrid-sd-tinyvae for SDXL | ||
| Download the model weight [Hybrid-sd-tinyvae-xl](https://huggingface.co/cqyan/hybrid-sd-tinyvae-xl) | ||
| ```bash | ||
| curl -L -O https://huggingface.co/cqyan/hybrid-sd-tinyvae-xl/blob/main/diffusion_pytorch_model.safetensors | ||
| ``` | ||
| T2I Comparison using one A100 GPU, The image order from left to right : SD1.4-VAE -> TAESD -> Hybrid-sd-tinyvae | ||
|  | ||
|  | ||
| Specify the model path using the `--taesd PATH` parameter. example: | ||
| ```bash | ||
| sd -m ../models/sd_xl_base_1.0.safetensors -p "a lovely cat" --taesd ../models/diffusion_pytorch_model.safetensors | ||
| ``` |