File size: 7,687 Bytes
f83ee2a 99c5814 f2882f0 99c5814 bf13c86 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 bf13c86 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 f2882f0 a0d6678 99c5814 bf13c86 f2882f0 99c5814 f83ee2a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 | <!doctype html>
<html lang="en">
<head>
<meta charset="utf-8" />
<meta name="viewport" content="width=device-width, initial-scale=1" />
<meta name="color-scheme" content="light dark" />
<title>Lernex</title>
<style>
* { box-sizing: border-box; }
body {
margin: 0;
background: transparent;
color: #d7dde8;
font-family: ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif;
letter-spacing: 0;
}
.org-card {
width: 100%;
max-width: 100%;
margin: 0;
padding: 0 0 8px;
}
.hero {
display: block;
width: 100%;
height: auto;
margin: 0 0 40px;
border-radius: 2px;
}
p {
margin: 0 0 18px;
color: #c6cedb;
font-size: clamp(18px, 2.1vw, 22px);
line-height: 1.55;
}
.intro {
color: #e7ebf2;
font-style: italic;
}
a {
color: inherit;
font-weight: 700;
text-decoration: underline;
text-underline-offset: 3px;
}
h2 {
margin: 52px 0 20px;
color: #eef2f8;
font-size: clamp(24px, 3vw, 30px);
line-height: 1.2;
font-weight: 800;
}
ul {
margin: 0 0 12px;
padding-left: 28px;
color: #c6cedb;
font-size: clamp(18px, 2.1vw, 22px);
line-height: 1.55;
}
li {
margin: 0 0 16px;
}
strong {
color: #eef2f8;
font-weight: 800;
}
.resources {
margin-bottom: 0;
}
@media (prefers-color-scheme: light) {
body { color: #202733; }
p, ul { color: #475467; }
.intro, h2, strong { color: #111827; }
}
@media (max-width: 700px) {
.hero { margin-bottom: 28px; }
h2 { margin-top: 36px; }
ul { padding-left: 22px; }
}
</style>
</head>
<body>
<main class="org-card">
<img
class="hero"
src="https://huggingface.co/spaces/Lernex/README/resolve/main/assets/lernex-org-banner.jpg"
alt="Lernex turns learning material into adaptive study paths and compact AI systems"
/>
<p class="intro">Welcome to the official Lernex organization on Hugging Face.</p>
<p>
<a href="https://www.lernex.net/" target="_blank" rel="noreferrer">Lernex</a>
is an AI-personalized learning platform built around a simple idea:
studying should adapt to the learner, the material, and the moment.
Lernex turns notes, PDFs, slides, classroom material, questions, and
learner goals into lessons, quizzes, flashcards, explanations, study
paths, feedback, and memory-aware tutoring.
</p>
<p>
This organization is where Lernex publishes the Metis model line and
related open research artifacts. Metis models are compact, efficient
language models built for the specific jobs inside a learning product:
tutoring, explanations, writing support, practice generation, concise
reasoning, and fast product loops that cannot depend on oversized
systems for every interaction.
</p>
<h2>What Lernex Is Building</h2>
<ul>
<li><strong>A continuous learning companion</strong>: Lernex is designed so a learner does not restart from zero every time they open a different class, generated lesson, quiz, flashcard set, or study path.</li>
<li><strong>Adaptive study material</strong>: source material can become structured lessons, checks for understanding, retrieval-friendly notes, and targeted practice.</li>
<li><strong>Personalized learning paths</strong>: Lernex uses learner goals, progress, confusion signals, and prior activity to shape what should come next.</li>
<li><strong>Fast learning workflows</strong>: learning tools should feel responsive enough for everyday studying, not like heavyweight one-off AI demos.</li>
</ul>
<h2>Why We Build Models</h2>
<p>
Education products need a different model shape than general chat
products. Some proprietary models are powerful but expensive or slow
for high-frequency learning loops. Some open models are affordable but
miss the tutoring, instruction-following, reasoning, or product-specific
behavior Lernex needs.
</p>
<p>
Metis is our answer to that mismatch: build compact models that are
useful where learning quality matters, efficient enough to run often,
and honest about the tradeoffs between size, reasoning, latency, and
cost.
</p>
<h2>Featured Models</h2>
<ul>
<li><strong><a href="https://huggingface.co/Lernex/Metis-1.4-base" target="_blank" rel="noreferrer">Metis-1.4 Base</a></strong> - current base release for the Metis-1.4 line.</li>
<li><strong><a href="https://huggingface.co/Lernex/Metis-1.4-chat" target="_blank" rel="noreferrer">Metis-1.4 Chat</a></strong> - conversational variant for tutoring, writing, explanations, and learning support.</li>
<li><strong><a href="https://huggingface.co/Lernex/Metis-1.4-think" target="_blank" rel="noreferrer">Metis-1.4 Think</a></strong> - reasoning-oriented variant for compact step-by-step problem solving.</li>
<li><strong><a href="https://huggingface.co/Lernex/Metis-1.3-base" target="_blank" rel="noreferrer">Metis-1.3 Base</a></strong>, <strong><a href="https://huggingface.co/Lernex/Metis-1.3-chat" target="_blank" rel="noreferrer">Chat</a></strong>, and <strong><a href="https://huggingface.co/Lernex/Metis-1.3-think" target="_blank" rel="noreferrer">Think</a></strong> - earlier hybrid Mamba-attention research releases.</li>
</ul>
<h2>Research Direction</h2>
<ul>
<li><strong>Compact tutoring models</strong>: smaller models that can still explain clearly, generate useful practice, and behave consistently in learning contexts.</li>
<li><strong>Efficient reasoning</strong>: model variants that spend compute where step-by-step problem solving matters instead of making every request expensive.</li>
<li><strong>Product-shaped evaluation</strong>: evaluating models on whether they help learners understand, practice, revise, and move forward, not only on broad benchmark prestige.</li>
<li><strong>Deployment practicality</strong>: keeping model size, serving cost, latency, and memory footprint in view from the start.</li>
</ul>
<h2>How This Connects to Lernex</h2>
<p>
Lernex uses AI to help learners bring messy material into a system that
can teach from it. The long-term goal is not just a better chatbot; it
is one learning relationship that carries context across generated
lessons, study paths, practice, feedback, and review.
</p>
<p>
The Metis line supports that goal by giving Lernex a model family it can
tune around real learning workflows instead of treating every AI job as
the same generic chat problem.
</p>
<h2>Resources</h2>
<ul class="resources">
<li>Website: <a href="https://www.lernex.net/" target="_blank" rel="noreferrer">lernex.net</a></li>
<li>Company page: <a href="https://www.lernex.net/company" target="_blank" rel="noreferrer">lernex.net/company</a></li>
<li>Product feedback: <a href="mailto:feedback@lernex.net">feedback@lernex.net</a></li>
<li>Business contact: <a href="mailto:business@lernex.net">business@lernex.net</a></li>
</ul>
</main>
</body>
</html>
|