Quickly create characters in JRPG style using Scenario's Anime Portraits Community Model
Step 1: Go to ChatGPT
Head over to ChatGPT and ask for descriptive keywords of the characters you wish to get as output.
Example:
Describe 5 different types of male and female JRPG game characters, how characters are dressed, and how they look, so that they are of different shapes (such as square, round, triangular). Make sure that each character has its own distinctive clothing color
Here are some ChatGPT outputs:
Dwarf: A male character with a short, stocky build with a square shape. He wears a brown leather tunic with fur trim, green pants, and black boots. He has a long, braided beard and a bald head, and carries a war hammer. His tunic has red accents.
Elf: A female character with a slender, graceful build with a triangular shape. She wears a green tunic with gold trim, a white skirt, and brown boots. She has long, pointed ears and flowing, white hair. She carries a longbow and a quiver of arrows. Her tunic has silver accents.
Ogre: A male character with a massive, hulking build with a round shape. He wears a black loincloth with red tribal markings, and metal gauntlets and boots. He has green skin and a bald head with a scar over one eye. He carries a giant club. His gauntlets and boots have gold accents.
Fairy: A female character who has a small, delicate build with a circular shape. She wears a pink and purple dress with a white apron and yellow stockings. She has delicate wings and long, blonde hair. She carries a wand with a star on top. Her dress has blue accents.
Golem: A male character who has a massive, solid build with a square shape. He wears a grey loincloth with silver runes, and metal gauntlets and boots. He has rocky skin and glowing blue eyes. He carries a massive stone hammer. His gauntlets and boots have red accents.
Modify the description as needed to include keywords or phrases separated by commas or ask chatgpt to do it for you. Add these keywords to the text description.
Final keywords result from example: Dwarf: A male character who has a short stocky build with a square shape. He wears a brown leather tunic with fur trim green pants and black boots. He has a long braided beard and a bald head and carries a war hammer. His tunic has red accents.
Step 2: Setup Model
On the left panel select 'Models'
Select Community Models: “Anime Portraits”
Copy a text prompt (the one you have generated from ChatGPT)
Paste it into the Prompt field
Setup Sampling Steps to 50 (during the generation of your images, try different options depending on the obtained results)
Insert a base picture in “image to image” field (AI gets a hint token that you want a neutral pose character model with no background. You can also generate variations without a base model to get more variation)
You can use these base character pictures with neutral poses: Link
Set and change Influence to 0 (you can experiment with the strength of Influence later on to get more options)
Change (in Advanced Options) Dimensions to 512x768 and Guidance to 20
You can change the Guidance, but remember that the higher the Guidance, the more the sharpness, and contrast will be increased - due to the fact that the AI tries to get closer to the model samples. The smaller the Guidance, the softer the lines of the drawing, and the structure of the drawing may collapse, due to the fact that the AI does not know what to focus on.
Step 3: Generate art
Press the button to Generate Images
Generate 4-8 images
Generate more images and pick the best-generated image for “Refine” and generate a new set of images. To do this, select the desired image - click on it and click on the settings button (three vertical lines) at the top left, after which a menu with the "Refine" button will be shown.
The following batches based on your generated images will understand more and more details from your Prompt.
From the generated images, you can make your own base pictures - from which you will start the next generation
If the Prompt does not give the desired result:
Separate the words with commas and change the weight of tokens with prompt expressions
You can use Negative Prompt to exclude keywords.
Change Prompt tokens to specify what exactly you want to see.
Change your Reference Image if needed.
Step 4: Download art
Download the right image, or good ideas or details. Note that you can make new base pictures from different parts using "image to image".
To quickly save images, use the Bulk Upload feature (maximum 100 images at a time). To do this, click the arrow button at the top right. An action menu will appear at the bottom with selected images. Click on the images you want to save, and then click on the Download button
You can download an image with automatic background removal. To do this, select the desired image - click on it and hit the settings button (three vertical lines) at the top left, after which a menu with the “Remove Background” button will appear.
Results
Time spent on the task:
ChatGPT keyword generation - 5 minutes
Scenario model setup - 2 minutes
Scenario 80 clicks on the "Generate Image" button (with 2 images to generate per batch, and copy-paste keywords, and change base pictures in “image to image”) - 20 minutes
- Download accepted images 100 Characters - 3 minutes
Select the final 5 characters:
All the characters were obtained without a single intervention of graphic editors!
Calculate how much time is saved with Scenario!
In about 30 minutes, you can get 5 characters (out of 160 variants received).
If you generate not 2 pictures for one click of the "Generate Pictures" button - but 8, you will get 640 variants.
Using the "Refine" function - you can deepen the search for variants in the right direction.
Examples
Prompt 1:
JRPG full-body character, neutral pose, Dwarf: A male character who has a short stocky build with a square shape. He wears a brown leather tunic with fur trim green pants and black boots. He has a long braided beard and a bald head and carries a war hammer. His tunic has red accents.
Prompt 2:
JRPG full-body character, neutral pose, Elf: A female character who has a slender graceful build with a triangular shape. She wears a green tunic with gold trim a white skirt and brown boots. She has long pointed ears and flowing white hair. She carries a longbow and a quiver of arrows. Her tunic has silver accents.
Prompt 3:
JRPG full-body character, neutral pose, Green skin Ogre: A male character who has a massive hulking build with a round shape. He wears a black loincloth with red tribal markings and metal gauntlets and boots. He has green skin and a bald head with a scar over one eye. He carries a giant club. His gauntlets and boots have gold accents.
Prompt 4:
JRPG full-body character, neutral pose, Golden skin Golem: A male character who has a massive solid build with a square shape. He wears a grey loincloth with silver runes and metal gauntlets and boots. He has rocky skin and glowing blue eyes. He carries a massive stone hammer. His gauntlets and boots have red accents.
