Update README.md
Browse files
README.md
CHANGED
|
@@ -5,6 +5,10 @@ license: creativeml-openrail-m
|
|
| 5 |
I preferred using DPM++ 2M SDE Karras for my model, but it's not necessarily the best
|
| 6 |
This 0.5.2 beta update is enhanced with 2.5d material. The model is still in beta so other than the surreal, artistic style and 3d, everything is still not quite right
|
| 7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
sdxl usage is completely different from all previous models:
|
| 9 |
|
| 10 |
1/ the number of cfg is greatly affected, can replace clip skip. if you want to create surreal images, cfg 3-4 is suitable, cartoon or 3d character images should be clear, cfg should be 5-6, and real people photos can sometimes be up to 7-8 but there are It can be lower, but only if you need high precision and detail to get cfg up to 9-10 or higher, I haven't tried it yet.
|
|
@@ -13,7 +17,9 @@ sdxl usage is completely different from all previous models:
|
|
| 13 |
|
| 14 |
3/ the way to get prompts will be different from other sd models, but I don’t know how, just feel that the prompts are too cumbersome and not as good as before and instead of writing in tag style, You can try to choose a descriptive writing style like Midjourney
|
| 15 |
|
| 16 |
-
4/
|
|
|
|
|
|
|
| 17 |
|
| 18 |
|
| 19 |
|
|
|
|
| 5 |
I preferred using DPM++ 2M SDE Karras for my model, but it's not necessarily the best
|
| 6 |
This 0.5.2 beta update is enhanced with 2.5d material. The model is still in beta so other than the surreal, artistic style and 3d, everything is still not quite right
|
| 7 |
|
| 8 |
+
Negative prompt:
|
| 9 |
+
|
| 10 |
+
(worst quality:1.5), (low quality:1.5), (normal quality:1.5), lowres, bad anatomy, bad hands, multiple eyebrow, (cropped), extra limb, missing limbs, deformed hands, long neck, long body, (bad hands), signature, username, artist name, conjoined fingers, deformed fingers, ugly eyes, imperfect eyes, skewed eyes, unnatural face, unnatural body, error, painting by bad-artist
|
| 11 |
+
|
| 12 |
sdxl usage is completely different from all previous models:
|
| 13 |
|
| 14 |
1/ the number of cfg is greatly affected, can replace clip skip. if you want to create surreal images, cfg 3-4 is suitable, cartoon or 3d character images should be clear, cfg should be 5-6, and real people photos can sometimes be up to 7-8 but there are It can be lower, but only if you need high precision and detail to get cfg up to 9-10 or higher, I haven't tried it yet.
|
|
|
|
| 17 |
|
| 18 |
3/ the way to get prompts will be different from other sd models, but I don’t know how, just feel that the prompts are too cumbersome and not as good as before and instead of writing in tag style, You can try to choose a descriptive writing style like Midjourney
|
| 19 |
|
| 20 |
+
4/ SDXL it seems that the more data, the less keywords are needed in prompt and negative prompt.
|
| 21 |
+
|
| 22 |
+
5/ That's all for now, I'll keep updating as new information or changes become available, anyway it's new technology for all of us.
|
| 23 |
|
| 24 |
|
| 25 |
|