Update README.md
Browse files
README.md
CHANGED
@@ -7,6 +7,22 @@ pipeline_tag: text-to-image
|
|
7 |
LoRA weights of Stable Diffusion XL for fast text-to-image generation.
|
8 |
|
9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
10 |
Note: Just find the normalCFG with 4-step is not working well. Trying to solve the issue.
|
11 |
|
12 |
[[paper](https://huggingface.co/papers/2405.18407)] [[arXiv](https://arxiv.org/abs/2405.18407)] [[code](https://github.com/G-U-N/Phased-Consistency-Model)] [[project page](https://g-u-n.github.io/projects/pcm)]
|
|
|
7 |
LoRA weights of Stable Diffusion XL for fast text-to-image generation.
|
8 |
|
9 |
|
10 |
+
Important Usage Guidance
|
11 |
+
Use DDIM or Euler instead of LCM for sampling! When using DDIM, set timestep_spacing="trailing".
|
12 |
+
|
13 |
+
The name of each LoRA weights indicates how many inference steps they should be applied.
|
14 |
+
|
15 |
+
The name of each LoRA weights indicates whether they are able to use normal CFGs or small CFGs
|
16 |
+
|
17 |
+
NormalCFG means that model equipped with the LoRA can use CFG value 2-9 for generation. Yet you should adjust the CFG values given the steps you applied.
|
18 |
+
When using fewer steps, you should use smaller CFGs. For example, use CFG 2.5 - 3.5 with 4 four steps and use CFG 3 - 6 with 8 steps. This is because that fewer-step means the model has fewer chance to fix the issues caused by the CFG.
|
19 |
+
|
20 |
+
SmallCFG means that the model equipped with the LoRA can use CFG value 1-2 for generation.
|
21 |
+
|
22 |
+
About the performance of normal CFG LoRAs.
|
23 |
+
|
24 |
+
|
25 |
+
|
26 |
Note: Just find the normalCFG with 4-step is not working well. Trying to solve the issue.
|
27 |
|
28 |
[[paper](https://huggingface.co/papers/2405.18407)] [[arXiv](https://arxiv.org/abs/2405.18407)] [[code](https://github.com/G-U-N/Phased-Consistency-Model)] [[project page](https://g-u-n.github.io/projects/pcm)]
|