Downloads · 30 days
16
9% of all-time downloads
Emerald-Wyvern-12B
Vortex5/Emerald-Wyvern-12B
Emerald-Wyvern-12B is a text generation model from Vortex5. Use it when you need the model to write or continue text. It is set up for transformers.
<style @import url('https://fonts.googleapis.com/css2?family=Cinzel+Decorative:wght@700&family=Uncial+Antiqua&family=MedievalSharp&display=swap');
All-time downloads
175
Public
Parameters
12.2B
24.5 GB on disk
Likes
5
Public
Repo makeup
Click a slice to open those files.
.safetensors24.5 GB · 100%
At a glance
- Task
- Text Generation
- Library
- transformers
- Model type
- mistral
- Access
- Public
- Created
- Nov 7, 2025
- Updated
- Nov 7, 2025
- SHA
- b638e0d0
Try a prompt
Base models
Model card
From the Hugging Face model README
<style>
@import url('https://fonts.googleapis.com/css2?family=Cinzel+Decorative:wght@700&family=Uncial+Antiqua&family=MedievalSharp&display=swap');
:root {
--bg-deep:#040705;
--bg-mid:#0a0f0a;
--bg-top:#111a12;
--emerald:#7fffb0;
--mint:#caffdc;
--gold:#f8e4a3;
--text-main:#eaf7ec;
--text-dim:#a7bfae;
}
*{box-sizing:border-box;}
.draconic-wrap {
position:relative;
margin:60px auto;
max-width:1100px;
padding:48px;
font-family:"MedievalSharp","Cinzel Decorative","Uncial Antiqua",serif;
color:var(--text-main);
background:radial-gradient(circle at 60% 20%,var(--bg-top),var(--bg-mid) 55%,var(--bg-deep) 100%);
border-radius:20px;
border:2px solid var(--emerald);
box-shadow:
0 0 15px rgba(60,90,60,0.3),
0 0 90px rgba(90,255,160,0.15),
inset 0 0 100px rgba(90,200,130,0.08);
}
.title-container {
text-align: center;
margin-bottom: 40px;
}
.title-main {
margin: 0;
font-size: 3.2rem;
letter-spacing: 3px;
text-transform: uppercase;
font-family: "Cinzel Decorative", serif;
color: var(--mint);
text-shadow:
0 0 6px rgba(120,255,180,0.7),
0 0 14px rgba(80,200,140,0.5),
0 0 25px rgba(100,255,180,0.4),
0 0 50px rgba(90,255,160,0.2);
background: none;
}
.subtitle {
margin-top: .6rem;
color: var(--text-dim);
font-style: italic;
text-shadow: 0 0 5px rgba(100,255,180,.3);
}
.hero{text-align:center;margin:30px auto;}
.hero img{
max-width:95%;
border-radius:14px;
border:2px solid var(--emerald);
box-shadow:
0 0 40px rgba(100,240,170,.25),
inset 0 0 8px rgba(80,200,130,.1);
}
.section{
margin:35px 0;
padding:28px 30px;
border-radius:14px;
background:linear-gradient(180deg,rgba(8,12,8,.92),rgba(6,9,6,.97));
border:1px solid rgba(90,220,150,.25);
box-shadow:
inset 0 0 25px rgba(0,0,0,.7),
0 0 40px rgba(90,220,150,.1);
}
.section-header{
padding-bottom:10px;
border-bottom:1px solid rgba(100,255,180,.15);
margin-bottom:10px;
}
.section-title{
font-family:"Uncial Antiqua",serif;
font-size:1.7rem;
letter-spacing:3px;
text-transform:uppercase;
color:var(--mint);
text-shadow:
0 0 8px rgba(150,255,200,.6),
0 0 18px rgba(100,240,160,.4);
}
.section-content p{
line-height:1.7;
color:var(--text-main);
text-shadow:0 0 5px rgba(80,150,100,.25);
}
details{
margin-top:14px;
border:1px solid rgba(120,250,180,.25);
border-radius:10px;
background:linear-gradient(180deg,rgba(15,20,15,.7),rgba(10,15,10,.85));
padding:10px 14px;
}
summary{
cursor:pointer;
font-family:"Cinzel Decorative",serif;
font-weight:600;
color:var(--mint);
text-shadow:0 0 8px rgba(120,255,180,.4);
list-style:none;
}
details summary::-webkit-details-marker{display:none;}
details > summary::before{
content:"▸";padding-right:6px;opacity:.8;
}
details[open] > summary::before{transform:rotate(90deg);}
pre{
background:#0a0f0a;
color:var(--text-main);
padding:16px 22px;
border-left:3px solid var(--emerald);
border-radius:8px;
overflow:auto;
font-size:13px;
box-shadow:inset 0 0 18px rgba(0,0,0,.7);
}
a{
color:var(--mint);
text-decoration:none;
border-bottom:1px dotted rgba(150,255,200,.4);
}
a:hover{
color:var(--gold);
text-shadow:0 0 8px rgba(200,255,220,.7);
}
ul{list-style:none;padding-left:10px;margin:0;}
</style>
<div class="draconic-wrap">
<div class="title-container">
<h1 class="title-main">Emerald Wyvern</h1>
</div>
<div class="hero">
<img src="https://cdn-uploads.huggingface.co/production/uploads/6669a3a617b838fda45637b8/2Md7RGyGnEaASRPyYVfba.png" alt="Shadowed-Wyvern"/>
</div>
<div class="section">
<div class="section-header"><h2 class="section-title">Overview</h2></div>
<div class="section-content">
<p><b>Emerald-Wyvern-12B</b> was made with a custom method, it merges —
<a href="https://huggingface.co/Vortex5/Luminous-Shadow-12B">Luminous-Shadow-12B</a>,
<a href="https://huggingface.co/Vortex5/Velvet-Orchid-12B">Velvet-Orchid-12B</a>, and
<a href="https://huggingface.co/Vortex5/Crimson-Twilight-12B">Crimson-Twilight-12B</a>,
using <a href="https://huggingface.co/Vortex5/Harmony-Bird-12B">Harmony-Bird-12B</a> as a base.
</p>
</div>
</div>
<div class="section">
<div class="section-header"><h2 class="section-title">Merge Configuration</h2></div>
<div class="section-content">
<details>
<summary>Expand YAML Configuration</summary>
<pre><code>
models:
- model: Vortex5/Luminous-Shadow-12B
parameters:
weight:
- filter: self_attn
value: [0.20, 0.35, 0.55, 0.75, 1.00, 0.90, 0.25, 0.25]
- value: 0.33
- model: Vortex5/Velvet-Orchid-12B
parameters:
weight:
- filter: mlp
value: [0.20, 0.30, 0.20, 0.7, 0.65, 0.66, 0.0, 0.30]
- value: 0.33
- model: Vortex5/Crimson-Twilight-12B
parameters:
weight:
- filter: self_attn
value: [0.25, 0.35, 0.45, 0.55, 0.60, 0.65, 0.88, 0.88]
- filter: mlp
value: [0.20, 0.30, 0.45, 0.65, 0.80, 0.90, 0.95, 0.85]
- value: 0.33
merge_method: flowsync
base_model: Vortex5/Harmony-Bird-12B
dtype: bfloat16
parameters:
strength: 1.0
balance: 0.5
tokenizer:
source: Vortex5/Harmony-Bird-12B
</code></pre>
</details>
</div>
</div>
<div class="section">
<div class="section-header"><h2 class="section-title">Custom Merge Method</h2></div>
<div class="section-content">
<details>
<summary>About</summary>
<h1 style="margin-top: 30px;">FlowSync</h1>
<p class="concept">
<strong>Concept:</strong> A coherence-weighted tensor merge algorithm that aligns donor model deltas
relative to a shared base, computes pairwise cosine similarities to measure inter-model agreement,
and aggregates them along the highest-consensus direction in parameter space. It reinforces semantically
consistent weight shifts while suppressing incoherent or noisy deviations using adaptive temperature scaling,
semantic gating (Nyström eigenvector extraction), and MAD-based normalization.
</p>
<h2>Key Parameters</h2>
<ul class="param-list">
<li><code>strength</code> (global): Scales how far the merged tensor moves from the base toward the consensus vector. Higher values emphasize donor influence.</li>
<li><code>balance</code> (global): Modulates bias between logic-dominant (attention/MLP) and style-dominant (embedding/output) layers.</li>
</ul>
</details>
</div>
</div>
<div class="section">
<div class="section-header"><h2 class="section-title">Credits</h2></div>
<div class="section-content">
<ul>
<li><strong><a href="https://huggingface.co/mradermacher" target="_blank">Team Mradermacher</a></strong> — Static & imatrix quants</li>
<li><strong><a href="https://huggingface.co/DeathGodlike" target="_blank">DeathGodlike</a></strong> — EXL3 quants</li>
<li><strong><a>Original creators and model authors</a></strong></li>
</ul>
</div>
</div>
</div>
- Task
- Text Generation
- Library
- transformers
- Type
- mistral
- Created
- Nov 7, 2025
- Updated
- Nov 7, 2025