bokyeong1015 commited on
Commit
ec6ef19
1 Parent(s): f728889

Update docs/description.md

Browse files
Files changed (1) hide show
  1. docs/description.md +2 -1
docs/description.md CHANGED
@@ -1,4 +1,4 @@
1
- This demo showcases a lightweight Stable Diffusion model (SDM) for general-purpose text-to-image synthesis. Our model **BK-SDM-Small** achieves **36% reduced** parameters and latency. This model is bulit with (i) removing several residual and attention blocks from the U-Net of SDM-v1.4 and (ii) distillation pretraining on only 0.22M LAION pairs (fewer than 0.1% of the full training set). Despite very limited training resources, our model can imitate the original SDM by benefiting from transferred knowledge.
2
 
3
  <center>
4
  <img alt="U-Net architectures and KD-based pretraining" img src="https://huggingface.co/spaces/nota-ai/compressed-stable-diffusion/resolve/91f349ab3b900cbfec20163edd6a312d1e8c8193/docs/fig_model.png" width="65%">
@@ -8,6 +8,7 @@ This demo showcases a lightweight Stable Diffusion model (SDM) for general-purpo
8
 
9
 
10
  ### Notice
 
11
  - This research was accepted to
12
  - [**ICML 2023 Workshop on Efficient Systems for Foundation Models** (ES-FoMo)](https://es-fomo.com/)
13
  - [**ICCV 2023 Demo Track**](https://iccv2023.thecvf.com/)
 
1
+ This demo showcases a lightweight Stable Diffusion model (SDM) for general-purpose text-to-image synthesis. Our model [**BK-SDM-Small**](https://huggingface.co/nota-ai/bk-sdm-small) achieves **36% reduced** parameters and latency. This model is bulit with (i) removing several residual and attention blocks from the U-Net of [SDM-v1.4](https://huggingface.co/CompVis/stable-diffusion-v1-4) and (ii) distillation pretraining on only 0.22M LAION pairs (fewer than 0.1% of the full training set). Despite very limited training resources, our model can imitate the original SDM by benefiting from transferred knowledge.
2
 
3
  <center>
4
  <img alt="U-Net architectures and KD-based pretraining" img src="https://huggingface.co/spaces/nota-ai/compressed-stable-diffusion/resolve/91f349ab3b900cbfec20163edd6a312d1e8c8193/docs/fig_model.png" width="65%">
 
8
 
9
 
10
  ### Notice
11
+ - The model weights are available at BK-SDM-{[Base](https://huggingface.co/nota-ai/bk-sdm-base), [Small](https://huggingface.co/nota-ai/bk-sdm-small), [Tiny](https://huggingface.co/nota-ai/bk-sdm-tiny)} and can be easily used with 🤗 Diffusers.
12
  - This research was accepted to
13
  - [**ICML 2023 Workshop on Efficient Systems for Foundation Models** (ES-FoMo)](https://es-fomo.com/)
14
  - [**ICCV 2023 Demo Track**](https://iccv2023.thecvf.com/)