Spaces:
Running
Running
File size: 14,750 Bytes
41a567b | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 | <!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Supra Mini v4 2M | SupraLabs Blog</title>
<style>
:root {
--bg: #0f0f0f;
--surface: #1a1a1a;
--border: #333;
--text: #e0e0e0;
--accent: #536bfe;
--muted: #888;
--font-mono: 'JetBrains Mono', 'Fira Code', monospace;
}
* { margin: 0; padding: 0; box-sizing: border-box; }
body {
background-color: var(--bg);
color: var(--text);
font-family: 'Inter', -apple-system, sans-serif;
line-height: 1.6;
padding: 2rem;
}
code, pre, .mono { font-family: var(--font-mono); }
.container { max-width: 900px; margin: 0 auto; }
header {
border-bottom: 2px solid var(--border);
padding-bottom: 2rem;
margin-bottom: 3rem;
display: flex;
justify-content: space-between;
align-items: flex-end;
}
.logo-area h1 {
font-size: 1.2rem;
text-transform: uppercase;
letter-spacing: 2px;
color: var(--accent);
line-height: 1;
display: flex;
align-items: center;
gap: 10px;
}
.logo-area a { text-decoration: none; color: inherit; }
.logo-area {
display: flex;
align-items: center;
gap: 10px;
font-weight: bold;
font-size: 1.2rem;
}
nav a {
color: var(--text);
text-decoration: none;
margin-left: 1.5rem;
font-size: 0.9rem;
border-bottom: 1px solid transparent;
}
nav a:hover { border-bottom: 1px solid var(--accent); }
.post-header { margin-bottom: 3rem; }
.post-header h2 {
font-size: 3rem;
line-height: 1.1;
margin-bottom: 1rem;
font-weight: 800;
}
.post-meta {
font-family: var(--font-mono);
color: var(--accent);
font-size: 0.9rem;
margin-bottom: 2rem;
}
.post-content {
background: var(--surface);
border: 1px solid var(--border);
padding: 3rem;
margin-bottom: 4rem;
}
.post-content h2 {
font-size: 1.8rem;
margin: 2.5rem 0 1rem 0;
color: var(--accent);
}
.post-content h2:first-child { margin-top: 0; }
.post-content p {
margin-bottom: 1.5rem;
font-size: 1.1rem;
color: var(--text);
}
.post-content ul {
margin-bottom: 1.5rem;
padding-left: 1.5rem;
}
.post-content li { margin-bottom: 0.5rem; font-size: 1.1rem; }
.post-content strong { color: #fff; }
.post-content code {
background: #111;
border: 1px solid var(--border);
padding: 2px 6px;
border-radius: 3px;
font-size: 0.95em;
color: var(--accent);
}
/* Code block */
.code-block {
background: #111;
border: 1px solid var(--border);
padding: 1.5rem;
margin: 2rem 0;
overflow-x: auto;
font-family: var(--font-mono);
font-size: 0.88rem;
line-height: 1.7;
color: #ccc;
}
.code-block .comment { color: var(--muted); }
.code-block .keyword { color: var(--accent); }
.callout {
border-left: 3px solid var(--accent);
background: #111;
padding: 1rem 1.5rem;
margin: 2rem 0;
font-family: var(--font-mono);
font-size: 0.95rem;
color: #ccc;
}
.callout span {
display: block;
color: var(--muted);
font-size: 0.8rem;
margin-bottom: 0.4rem;
}
/* Output example box */
.output-example {
border: 1px solid var(--border);
background: #111;
padding: 1.5rem;
margin: 1.5rem 0;
}
.output-example .prompt-label {
font-family: var(--font-mono);
font-size: 0.75rem;
color: var(--accent);
margin-bottom: 0.4rem;
}
.output-example .prompt-text {
font-weight: 700;
color: #fff;
margin-bottom: 1rem;
font-size: 1rem;
}
.output-example .output-label {
font-family: var(--font-mono);
font-size: 0.75rem;
color: var(--muted);
margin-bottom: 0.4rem;
}
.output-example .output-text {
color: var(--text);
font-size: 0.95rem;
font-style: italic;
line-height: 1.7;
}
.table-wrap { overflow-x: auto; margin: 2rem 0; }
table {
width: 100%;
border-collapse: collapse;
font-family: var(--font-mono);
font-size: 0.9rem;
}
th {
background: #111;
color: var(--accent);
padding: 0.75rem 1rem;
text-align: left;
border: 1px solid var(--border);
}
td {
padding: 0.7rem 1rem;
border: 1px solid var(--border);
color: var(--text);
}
tr:nth-child(even) td { background: #111; }
.tags { display: flex; gap: 0.5rem; margin-top: 2rem; flex-wrap: wrap; }
.tag {
font-family: var(--font-mono);
font-size: 0.7rem;
padding: 2px 8px;
border: 1px solid var(--border);
border-radius: 4px;
color: var(--muted);
}
footer {
margin-top: 6rem;
padding-bottom: 2rem;
font-size: 0.8rem;
color: var(--muted);
text-align: center;
}
@media (max-width: 600px) {
.post-header h2 { font-size: 2rem; }
.post-content { padding: 1.5rem; }
header { flex-direction: column; align-items: flex-start; gap: 1rem; }
nav a { margin-left: 0; margin-right: 1rem; }
}
</style>
</head>
<body>
<div class="container">
<header>
<div class="logo-area" style="font-size: 1.5em;">
<a href="./index.html"><h1><img src="./image.png" style="height: 2em"> SupraLabs_</h1></a>
</div>
<nav>
<a href="./index.html#news">News</a>
<a href="https://huggingface.co/SupraLabs" target="blank">HuggingFace</a>
<a href="./index.html#hardware">Hardware</a>
</nav>
</header>
<article>
<div class="post-header">
<div class="post-meta">// 2026-05-15 | Release</div>
<h2>🦅 Supra Mini v4 2M<br>is here.</h2>
</div>
<div class="post-content">
<p>Today we are releasing <strong>Supra Mini v4 2M:</strong> the fourth version of our Supra Mini series and our biggest leap yet. Trained on <strong>3 billion tokens</strong> of Fineweb-Edu for 2 epochs, v4 pushes our parameter count to 2.6M while keeping the model light enough to run on any CPU.</p>
<h2>What changed from v3?</h2>
<p>Look at the numbers: <strong>v4 has ~5× more parameters than v3</strong>. We went from 467k to 2.6M parameters. This is not just a bigger model, the entire config was rethought to fit more capacity while keeping the architecture clean and the training fast.</p>
<div class="callout">
<span>// supra mini v4 2m — model config</span>
Parameters → 2,623,104 (~2.6M)<br>
Architecture → Llama<br>
Vocab size → 8,192 (custom BPE)<br>
Hidden size → 128<br>
Intermediate → 512<br>
Layers → 6<br>
Attention heads → 4<br>
Context length → 1,024 tokens<br>
Trained in → bfloat16
</div>
<h2>Training setup</h2>
<p>We trained v4 on a <strong>single NVIDIA RTX 5060 Ti 16GB</strong> in approximately 3 hours for 2 epochs. The dataset is the first 3 billion tokens of Sample-10BT from <strong>Fineweb-Edu</strong>, streamed and tokenized on the fly with our custom BPE tokenizer.</p>
<p>The final training loss after 2 epochs came in at <strong>4.618</strong>. The full training code, tokenizer, training loop, and inference script, is available directly in the model repo.</p>
<h2>Benchmarks</h2>
<p>We evaluated v4 using <code>lm-eval</code> on three tasks. The random baselines are included so you can judge fairly.</p>
<div class="table-wrap">
<table>
<thead>
<tr>
<th>Task</th>
<th>Score</th>
<th>Random baseline</th>
<th>Delta</th>
</tr>
</thead>
<tbody>
<tr><td>ARC_Easy</td><td>0.3152</td><td>0.25 (25%)</td><td>+6.5pp above random</td></tr>
<tr><td>Wikitext (PPL)</td><td>3.1652</td><td>—</td><td>lower is better</td></tr>
<tr><td>BLiMP</td><td>0.607</td><td>0.50 (50%)</td><td>+10.7pp above random</td></tr>
</tbody>
</table>
</div>
<p>A 2.6M parameter model beating random by over 10 points on BLiMP, a test of grammatical knowledge, is a solid result at this scale. Not GPT-4, obviously, but that is never the point with Supra Mini.</p>
<h2>Example outputs</h2>
<p>Here is what v4 generates at <code>temperature=0.5</code>, <code>top_k=25</code>, <code>top_p=0.9</code>:</p>
<div class="output-example">
<div class="prompt-label">// prompt</div>
<div class="prompt-text">"Artificial intelligence is "</div>
<div class="output-label">// output</div>
<div class="output-text">"Artificial intelligence is the idea of the theory that the world has a very high-performance technology, which is also more important to society's lives than people who are being able to find their own knowledge and understanding how it can be used for future generations..."</div>
</div>
<div class="output-example">
<div class="prompt-label">// prompt</div>
<div class="prompt-text">"The main concept of physics is "</div>
<div class="output-label">// output</div>
<div class="output-text">"The main concept of physics is 'animi-hisi', and therefore the universe's own light. In this case, a theory that is not only used to explain what it can be called 'the universe' or 'two planets, which are exactly about the earth's gravitational energy, but also in reality, we know how much things do..."</div>
</div>
<p>The model clearly has a coherent sense of topic, it stays on subject and builds sentences. It hallucinates and drifts (as all base models at this scale do), but the fluency is real.</p>
<h2>How to run it</h2>
<p>Drop this into any Python environment with Transformers installed:</p>
<div class="code-block">
<span class="keyword">from</span> transformers <span class="keyword">import</span> pipeline<br>
<span class="keyword">import</span> torch<br>
<br>
pipe = pipeline(<br>
<span class="string">"text-generation"</span>,<br>
model=<span class="string">"SupraLabs/Supra-Mini-v4-2M"</span>,<br>
device_map=<span class="string">"auto"</span>,<br>
torch_dtype=torch.float16 <span class="keyword">if</span> torch.cuda.is_available() <span class="keyword">else</span> torch.float32<br>
)<br>
<br>
result = pipe(<br>
<span class="string">"The importance of education is"</span>,<br>
max_new_tokens=150,<br>
do_sample=True,<br>
temperature=0.5,<br>
top_k=25,<br>
top_p=0.9,<br>
repetition_penalty=1.2<br>
)<br>
<span class="keyword">print</span>(result[0][<span class="string">'generated_text'</span>])
</div>
<h2>What's next?</h2>
<p>v4 is a base model, it is not fine-tuned for instruction following or chat. The next experiments on our roadmap include fine-tuning on instruction datasets, exploring quantization at this new scale, and continuing to push the parameter count while keeping training accessible to everyone with a consumer GPU.</p>
<p><strong>The model is live on HuggingFace. Go try it.</strong></p>
<div class="callout">
<span>// links</span>
Model → huggingface.co/SupraLabs/Supra-Mini-v4-2M<br>
License → Apache 2.0<br>
Series → Supra Mini collection on HuggingFace
</div>
<div class="tags">
<span class="tag">#release</span>
<span class="tag">#supra-mini-v4</span>
<span class="tag">#tinyml</span>
<span class="tag">#llama</span>
<span class="tag">#open-source</span>
<span class="tag">#fineweb-edu</span>
<span class="tag">#edge-ai</span>
</div>
</div>
</article>
<footer>
<p class="mono">© 2026 SupraLabs // Built for the community.</p>
</footer>
</div>
</body>
</html> |