Downloads · 30 days
64
4% of all-time downloads
CardProjector-27B-v4
SlerpE/CardProjector-27B-v4
CardProjector-27B-v4 is a image-text-to-text model from SlerpE. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. The card lists the license as gemma.
<!DOCTYPE html <html lang="en" <head <meta charset="UTF-8" <titleCardProjector-v4 - Character Generation Model</title <link rel="preconnect" href="https://fonts.googleapis.com" <link rel="preconnect" href="https://fon…
All-time downloads
1.6K
Public
Parameters
27.4B
54.9 GB on disk
Likes
35
Public
Repo makeup
Click a slice to open those files.
.safetensors54.9 GB · 100%
At a glance
- Task
- Image-Text-to-Text
- License
- gemma
- Model type
- gemma3
- Access
- Public
- Created
- Apr 10, 2025
- Updated
- Jun 28, 2025
- SHA
- b43d4cc8
Try a prompt
Base models
Model card
From the Hugging Face model README
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<title>CardProjector-v4 - Character Generation Model</title>
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Philosopher:ital,wght@0,400;0,700;1,400;1,700&display=swap" rel="stylesheet">
<style>
body {
font-family: 'Philosopher', 'Segoe UI', Arial, sans-serif;
line-height: 1.6;
margin: 0;
padding: 0;
/* --- Deep Teal/Blue Night Sky Background --- */
background: #0a1829; /* Dark Navy Blue base */
}
.container {
width: 100%;
margin: 0;
/* --- Night Sky Gradient from Image (Teal/Cyan focus) --- */
background: linear-gradient(135deg, #0a1829 0%, #0d2c3b 50%, #0e3d4a 100%); /* Dark Navy -> Dark Teal -> Tealish Blue */
padding: 30px;
box-sizing: border-box;
/* --- Text Color - Light Grey (Pony's Body) --- */
color: #d8d8e0; /* Kept light grey */
/* --- Border & Glow - Softer Gold (Mane Highlight) --- */
border: 3px solid #F0E68C; /* Khaki (Softer Gold) */
box-shadow: 0 0 15px 5px rgba(240, 230, 140, 0.4); /* Softer gold glow */
}
h1, h2 {
/* --- Headings - Softer Gold (Mane Mid-tone) --- */
color: #EEE8AA; /* PaleGoldenrod */
/* --- Border - Muted Gold/Brown (Mane Shadow) --- */
border-bottom: 2px solid #BDB76B; /* DarkKhaki */
padding-bottom: 5px;
margin-top: 0;
text-shadow: 1px 1px 2px rgba(0, 0, 0, 0.7); /* Shadow for readability */
}
/* --- Links - Pale Yellow/Gold (Mane Highlight) --- */
a {
color: #FFFACD; /* LemonChiffon */
text-decoration: none;
}
a:hover {
color: #ffffff; /* White on hover */
text-decoration: underline;
}
/* Specific Link */
p a[href*="WoonaAI"] {
color: #FFFACD; /* LemonChiffon */
}
.model-image {
text-align: center;
margin: 25px auto;
max-width: 800px;
}
.model-image img {
width: 100%;
height: auto;
border-radius: 10px;
/* --- Image Border & Glow - Softer Gold --- */
border: 3px solid #F0E68C; /* Khaki */
box-shadow: 0 0 15px 5px rgba(240, 230, 140, 0.4); /* Softer gold glow */
display: block;
}
/* --- Table Styles - Night Theme with Soft Gold/Grey --- */
table {
width: 100%;
border-collapse: collapse;
margin: 25px 0;
background: none;
border-radius: 0;
overflow: visible;
border: none;
}
th, td {
padding: 14px 18px;
text-align: left;
/* --- Table Text - Light Grey --- */
color: #d8d8e0;
border-bottom: none;
}
tr:last-child td {
border-bottom: none;
}
th {
background-color: transparent;
font-weight: bold;
/* --- Table Headers - Softer Gold --- */
color: #EEE8AA; /* PaleGoldenrod */
text-align: left;
border-bottom: none;
}
tr:hover {
background: none;
}
/* Emphasized Table Text - Pale Yellow/Gold */
td strong, td u {
color: #FFFACD; /* LemonChiffon */
}
/* --- Highlight/Warning Blocks - Teal/Grey Accents --- */
.highlight, .warning {
padding: 15px 20px;
margin: 20px 0;
border-radius: 5px;
border-left-width: 5px;
border-left-style: solid;
color: #f0f0f8; /* Keep slightly lighter text */
}
.highlight {
/* --- Highlight Block - Teal Accent (Sky) --- */
background: rgba(32, 178, 170, 0.15); /* Semi-transparent LightSeaGreen */
border-left-color: #20B2AA; /* LightSeaGreen */
}
.highlight h3 {
/* --- Highlight Heading - Brighter Teal --- */
color: #48D1CC; /* MediumTurquoise */
margin-top: 0;
}
.warning {
/* --- Warning Block - Grey Accent (Pony Body) --- */
background: rgba(119, 136, 153, 0.2); /* Semi-transparent LightSlateGray */
border-left-color: #778899; /* LightSlateGray */
color: #e8e8f8;
padding: 15px 20px;
margin: 20px 0;
border-radius: 5px;
border-left-width: 5px;
border-left-style: solid;
}
.warning h3 {
color: #A0AEC1; /* Lighter Slate Gray */
margin-top: 0;
}
/* --- Badges - Dark Grey (Pony Body) & Pale Gold Text --- */
.badge {
display: inline-block;
padding: 6px 12px;
border-radius: 4px;
/* --- Badge Background - Dark Grey --- */
background: #696969; /* DimGray */
/* --- Badge Text - Pale Yellow/Gold --- */
color: #FFFACD; /* LemonChiffon */
font-weight: bold;
margin: 5px;
box-shadow: 1px 1px 3px rgba(0, 0, 0, 0.6);
}
footer {
margin-top: 30px;
text-align: center;
opacity: 0.85;
/* --- Footer Text - Light Grey --- */
color: #d8d8e0;
text-shadow: 1px 1px 2px rgba(0, 0, 0, 0.7);
}
/* Chat Example Styling - Teal/Gold Theme */
.chat-example {
border-radius: 12px;
padding: 20px;
margin: 25px 0;
/* --- Chat Box Border - Dark Teal/Grey --- */
border: 1px solid #2F4F4F; /* DarkSlateGray */
}
.message {
margin: 15px 0;
padding: 15px;
border-radius: 8px;
max-width: 80%;
}
.message.user {
background: #1A2E38; /* Darker Blue-Teal */
/* --- User Message Border - Muted Gold --- */
border: 1px solid #BDB76B; /* DarkKhaki */
margin-left: auto;
}
.message.assistant {
background: #2A2A3A; /* Dark Grey-Blue */
/* --- Assistant Message Border - Grey --- */
border: 1px solid #778899; /* LightSlateGray */
margin-right: auto;
}
.role {
/* --- Role Text - Pale Yellow/Gold --- */
color: #FFFACD; /* LemonChiffon */
font-weight: bold;
margin-bottom: 8px;
font-size: 0.9em;
}
pre {
/* --- Code Block Background - Very Dark Navy --- */
background: #0a1829;
padding: 12px;
border-radius: 6px;
white-space: pre-wrap;
margin: 10px 0;
/* --- Code Text - Pale Yellow/Gold --- */
color: #FFFACD; /* LemonChiffon */
/* --- Code Border - Semi-Transparent Soft Gold --- */
border: 1px solid rgba(240, 230, 140, 0.3);
font-family: Consolas, Monaco, 'Andale Mono', 'Ubuntu Mono', monospace;
}
</style>
</head>
<body>
<!-- Весь контент остается внутри .container -->
<div class="container">
<h1>🎨 CardProjector-v4</h1> <!-- Title from second card -->
<p><a href="https://huggingface.co/WoonaAI">WoonaAi</a> presents...</p> <!-- Added author link -->
<div class="model-image">
<!-- Image from second card -->
<img src="https://cdn-uploads.huggingface.co/production/uploads/6336c5b3e3ac69e6a90581da/O2G9d-EhzpC1dNXA-6UzH.png" alt="CardProjector Model Visualization">
</div>
<h2>🚀 Overview</h2> <!-- Heading from second card -->
<!-- Overview content from second card -->
<p>CardProjector is a specialized series of language models, fine-tuned to generate character cards for <strong>SillyTavern</strong> and <strong>for creating characters in general</strong>. These models are designed to assist creators and roleplayers by automating the process of crafting detailed and well-structured character cards, ensuring compatibility with SillyTavern's format.</p>
<h2>🎨 CardProjector v4 tricks</h2> <!-- Heading changed, emoji kept for style consistency -->
<!-- List content from second card's "What's new" section -->
<ul>
<li>Absolute focus on personality development! This version places an absolute emphasis on designing character personalities, focusing on depth and realism. Eight (!) large datasets were collected, oriented towards all aspects of in-depth personality development. Extensive training was also conducted on a dataset of MBTI profiles with Enneagrams from psychology. The model was carefully trained to select the correct personality type according to both the MBTI and Enneagram systems. I highly recommend using these systems (see Usage recommendations); they provide an incredible boost to character realism. I conducted numerous tests with many RP models ranging from 24-70B parameters, and the MBTI profile system significantly impacts the understanding of the character's personality (especially on 70B models), making the role-playing performance much more realistic. You can see an example of a character's MBTI profile <a href="https://www.personality-database.com/profile/7610/muffins-derpy-hooves-ditzy-doo-my-little-pony-friendship-is-magic-2010-mbti-personality-type">here</a>. Currently, version V4 yields the deepest and most realistic characters.</li>
<li>Reduced likelihood of positive bias! I collected a large toxic dataset focused on creating and editing aggressive, extremely cruel, and hypersexualized characters, as well as transforming already "good harmless" characters into extremely cruel anti-versions of the original. Thanks to this, it was possible to significantly reduce the overall positive bias (especially in Gemma 3, where it is quite pronounced in its vanilla state), and make the model more balanced and realistic in terms of creating negative characters. It will no longer strive at all costs to create a cute, kind, ideal character, unless specifically asked to do so. All you need to do is just ask the model to "not make a positive character, but create a realistic one," and with that one phrase, the entire positive bias goes away.</li>
<li>Moving to Gemma 3! After a series of experiments, it turned out that this model is ideally suited for the task of character design, as it possesses much more developed creative writing skills and higher general knowledge compared to Mistral 2501 in its vanilla state. Gemma 3 also seemed much more logical than its French competitor.</li>
<li>Vision ability! Due to the reason mentioned in the point above, you can freely use vision in this version. If you are using GGUF, you can download the mmproj model for the 27B version from bartowski (a vanilla mmproj will suffice, as I didn't perform vision tuning).</li>
<li>The overall quality of character generation has been significantly increased by expanding the dataset approximately 5 times compared to version V3.</li>
<li>This model is EXTREMELY sensitive to the user's prompt. So you should give instructions with caution, carefully considering.</li>
<li>In version V4, I concentrated only on one model size, 27B. Unfortunately, training multiple models at once is extremely expensive and consumes too much effort and time, so I decided it would be better to direct all my resources into just one model to avoid scattering focus. I hope you understand 🙏</li>
</ul>
<div class="highlight">
<h3>💡 Usage Recommendations</h3> <!-- Heading from second card -->
<!-- Content from second card's "Usage recommendations" -->
<p><strong>Chat Template</strong>: Gemma</p>
<p><strong>Balanced output</strong>:<br>
Temperature: 0.8<br>
Top-P: 0.92<br>
Rp.Pen: 1.07<br>
Top-K: 100<br>
Rep Pen Range: 360<br>
Rep Pen Slope: 0.7<br></p>
<p><strong>The character creation process recomendations (updated)</strong>: In version V4, I want to recommend a new usage method. I strongly recommend generating characters directly in YAML format, rather than starting with natural text first and then converting it to YAML. Gemma 3 turned out to be much better for this use case; you immediately get ready-made, structured characters described in significant detail (in V3, I advised starting the design specifically with natural text, because Mistral and especially Qwen 2.5 had the problem of overly brief character descriptions when generating them directly in YAML).</p>
<p><strong>VERY IMPORTANT THING</strong>: As I mentioned earlier, this model is EXTREMELY sensitive to the prompt. When I say EXTREMELY, I really mean EXTREMELY. On one hand, this is a positive aspect—the model does exactly what you tell it to. On the other hand, you need to be very careful about the instructions you give it. You don’t necessarily have to write an overly detailed prompt—just include everything you need to get from the card. Be concise but thorough. The model won’t assume anything extra on behalf of the user—only what the user explicitly asks for. For example, if you want a card with a deeply developed personality, you should specify that in the prompt, or if you want to focus on appearance attributes, you need to emphasize that. That’s how CardProjector V4 works.</p>
<p><strong>MBTI personality profilies: </strong> Absolutely must have! This improves the character's personalities several times over the usual description. I strongly recommend using MBTI personality profiles as an add-on to the character card (see the tricks section regarding the MBTI system): first, generate the character card, for example, in YAML format, and when your character is ready, simply ask the model to supplement this work with an MBTI profile and Enneagram, placing it in a separate section within the card. Generally, you can ask the model to create the personality as an MBTI profile right from the first iteration, when you're just starting to generate the character. It's just that the method with the add-on seemed more convenient to me—there's no fundamental difference here. </p>
</div>
<!-- Add this section after the Usage Recommendations and before the Content Notice -->
</details>
<h2>💬 Examples</h2>
<details class="chat-example">
<summary>
<h2>MBTI usage (Let's imagine that we've already worked with a character in a chat before)</h2>
</summary>
<div class="message user">
<div class="role">User</div>
<div class="content">Please create an MBTI profile for my character in YAML format.
</div>
</div>
<div class="message assistant">
<div class="role">CardProjector</div>
<div class="content">
<pre>character_name: Mythic Horizon
profile: personality_type: ENFP - 3w2 - sx/sp - Sanguine [Dominant] - Scorpio Sun mbti_letter_stats: - type: E percentage: 100% interpretation: "Mythic is highly extroverted, thriving on social interaction and external stimulation. He draws energy from engaging with others and enjoys being the center of attention." - type: N percentage: 85% interpretation: "He leans heavily towards intuition, preferring abstract thinking and imaginative ideas over concrete details. He often sees possibilities where others see limitations." - type: F percentage: 97% interpretation: "Mythic is deeply feeling, making decisions based on values and empathy rather than logic alone. He prioritizes emotional harmony in both personal and professional relationships." - type: P percentage: 82% interpretation: "He’s flexible and adaptable, preferring to keep his options open rather than adhering to strict plans. This spontaneity adds an element of unpredictability to his actions."
enneagram_type: 3w2 - The Achiever with a Performer wing enneagram_interpretation: | "Mythic is driven by a desire to succeed and be recognized for his accomplishments. As a 3w2, he combines ambition with a strong need to connect with others and gain their approval. His core fear is being perceived as worthless or incapable, which fuels his relentless pursuit of success. Conversely, his core desire is to feel valuable and worthwhile, motivating him to excel in all areas of life."
cognitive_functions: - Ne (Extraverted Intuition): "His dominant function allows him to generate creative ideas and explore multiple possibilities simultaneously. He thrives on brainstorming and envisioning future scenarios." - Fi (Introverted Feeling): "As his auxiliary function, Fi guides his decision-making process based on personal values and emotions. He strives to align his actions with his inner moral compass." - Te (Extraverted Thinking): "Though not as prominent, Te provides Mythic with the ability to organize and execute plans efficiently when necessary. It surfaces in moments where practicality takes precedence over idealism." - Si (Introverted Sensing): "This inferior function manifests in occasional nostalgia or a tendency to draw on past experiences for guidance. However, it’s less developed compared to his other functions."
traits: positive: - Charismatic: "Mythic has a magnetic personality that draws people towards him. His charm makes him a natural leader and social magnet." - Creative: "He possesses a vivid imagination and a knack for thinking outside the box. His creativity shines through in both artistic endeavors and problem-solving." - Empathetic: "Mythic is deeply attuned to the emotions of others, often going out of his way to offer support and understanding." - Adaptable: "His flexible nature allows him to thrive in dynamic environments, easily adjusting to new circumstances and challenges." - Ambitious: "Driven by a desire to succeed, Mythic sets high goals for himself and works tirelessly to achieve them." negative: - Impulsive: "His spontaneity can sometimes lead to rash decisions without considering long-term consequences." - Inconsistent: "Mythic’s preference for keeping options open can make him unreliable or unpredictable in certain situations." - Image-Conscious: "He places significant importance on how others perceive him, which can lead to vanity or a need for constant validation." - Overwhelmed by Routine: "Mythic struggles with monotony and repetitive tasks, often seeking novelty even when stability would be more beneficial." - Emotionally Volatile: "Despite his empathy, Mythic’s intense emotional responses can sometimes overwhelm him or those around him."
interests: - Fashion and Aesthetics: "Mythic has a keen eye for style and enjoys experimenting with bold, unconventional looks that reflect his unique personality." - Performing Arts: "Whether acting, dancing, or singing, he finds joy in creative expression and loves being on stage." - Socializing: "He thrives in social settings, whether hosting parties, attending events, or simply engaging in deep conversations with friends." - Travel and Exploration: "Mythic is drawn to new experiences and cultures, always eager to explore unfamiliar territories." - Self-Improvement: "He’s constantly seeking ways to grow personally and professionally, whether through books, workshops, or mentorship opportunities."
quirks: - Always carries a small notebook to jot down ideas or inspirations as they come to him. - Has a habit of tapping his fingers rhythmically when deep in thought or feeling anxious. - Prefers to sleep with the curtains open to let in moonlight, finding it calming and inspiring. - Often hums or sings softly to himself while working, especially during creative tasks. - Collects vintage trinkets and accessories, each with its own story and sentimental value. </pre> </div> </div>
</details> <!-- Example sections from the second card are omitted as they were empty placeholders --> <!-- Safety section from the second card is replaced by the original warning, which covers similar concerns --> <div class="warning"> <h3>⚠️ Safety</h3> <!-- Kept original heading --> <!-- Kept original warning text, relevant due to tags/example warning in second card --> <p>This model learned on cards for Silly Tavern. I think comments are unnecessary here...</p> </div> <h2>🔧 Technical Specs</h2> <!-- Kept original heading --> <!-- Badges updated with info from second card's metadata --> <p> <span class="badge">Base Model: google/gemma-3-27b-it</span> <span class="badge">License: gemma</span> <span class="badge">Language: English</span> </p> <footer> <p>© 2024 WoonaAI | 🤗 Hosted on Hugging Face</p> <!-- Kept original footer --> </footer> </div> <!-- Конец .container --> </body> </html>- Task
- Image-Text-to-Text
- Type
- gemma3
- License
- gemma
- Languages
- en
- Featherless Ai
- live
- Created
- Apr 10, 2025
- Updated
- Jun 28, 2025