- Jun 9, 2019
- 344
- 1,336
- Jun 9, 2019
- 344
- 1,336
You don't have permission to view the spoiler content.
Log in or register now.
- Nov 4, 2019
- 569
- 726
You don't have permission to view the spoiler content.
Log in or register now.
You don't have permission to view the spoiler content.
Log in or register now.
maybe there are coders among us? Let's start writing the continuation of the game ourselves, judging by the art it should turn out well
- Nov 4, 2019
- 569
- 726
- Jun 9, 2019
- 344
- 1,336
- Jun 19, 2018
- 20
- 22
A direct continuation to this scene:
I tryed to train LoRA, with tiny dataset (10 images), on my PC, and failed. Obviously because it's a first try).
For training i use this program:
For training i use this program:
You must be registered to see the links
- Jun 19, 2018
- 20
- 22
Well, for LoRA from just 10 barely tagged images, i get pretty good results.
So 10 images and bad tagging can be enough for character and style)
You don't have permission to view the spoiler content.
Log in or register now.
You don't have permission to view the spoiler content.
Log in or register now.
So 10 images and bad tagging can be enough for character and style)
You don't have permission to view the spoiler content.
Log in or register now.
Last edited:
I used reference images, and "Reference" module from ControlNet with "reference_only" preprocessor.
Workflow is similar to "Depth".
You don't have permission to view the spoiler content.
Log in or register now.
- Mar 30, 2019
- 86
- 100
I could offer a bit of RenPy coding experience (basic) and a working SD_Forge Setup with a decent GPU.
- Nov 4, 2019
- 569
- 726
I'm still learning how to make AI art but my comp is pretty decent.
Ryzen 7 3800X 8-core processor
32gb of RAM (3200mhz)
2T M.2 card
RTX 3060 (12gb VRAM)
want sabia to get raped. its been almost a decade since she last got it
- Jul 12, 2017
- 1,154
- 1,104
Go back to page one if you want to make it happen
I've been dressing Sabia and Neve with naked apron
- Jun 9, 2019
- 344
- 1,336
- Sep 18, 2018
- 10
- 47
Experienced users have known this for a long time, but to the newbies I say:
Just like you customize Lora's slider to define Lora's weight, you can customize ANY word/definition to get unusual (or whatever you want) results.
I'll show you this with an example of one image in the Civitai gallery. Here is the result of a common query (pov male):
Now I'm adding a side view tag. But it doesn't have any effect:
If I leave "Side View" and remove "POV", however, I get this result:
However, we can weight the "Side View" higher to make it more meaningful.
To do this, we need to bracket the definition, enter a colon and the desired weight. 1 is the base weight. For example, I want to give "Side view" a weight of 1.4. by raising it. Then I write (side view:1.4). I enter "Pov" and "(side view:1.4)" and these are the results I get:
See? We got both an eye view and at the same time, a slightly side view.
Just like you customize Lora's slider to define Lora's weight, you can customize ANY word/definition to get unusual (or whatever you want) results.
I'll show you this with an example of one image in the Civitai gallery. Here is the result of a common query (pov male):
You don't have permission to view the spoiler content.
Log in or register now.
Now I'm adding a side view tag. But it doesn't have any effect:
You don't have permission to view the spoiler content.
Log in or register now.
If I leave "Side View" and remove "POV", however, I get this result:
You don't have permission to view the spoiler content.
Log in or register now.
However, we can weight the "Side View" higher to make it more meaningful.
To do this, we need to bracket the definition, enter a colon and the desired weight. 1 is the base weight. For example, I want to give "Side view" a weight of 1.4. by raising it. Then I write (side view:1.4). I enter "Pov" and "(side view:1.4)" and these are the results I get:
You don't have permission to view the spoiler content.
Log in or register now.
See? We got both an eye view and at the same time, a slightly side view.
- Jun 9, 2019
- 344
- 1,336
