Repository navigation
Expand file tree
/
Copy pathlived-lora-studio.html
More file actions
125 lines (120 loc) · 9.78 KB
/
Copy pathlived-lora-studio.html
File metadata and controls
125 lines (120 loc) · 9.78 KB
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
<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta http-equiv="Content-Security-Policy" content="default-src 'self'; script-src 'self' https://static.cloudflareinsights.com; style-src 'self'; img-src 'self' data:; font-src 'self'; connect-src 'self' https://cloudflareinsights.com; media-src 'self'; object-src 'none'; base-uri 'self'; form-action 'self' mailto:; upgrade-insecure-requests">
<meta name="referrer" content="strict-origin-when-cross-origin">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>liveD LoRA Studio — Reviewed Data and NVIDIA GPU Selection | PhantomBugz</title>
<meta name="description" content="liveD LoRA Studio source preview includes encrypted review, approved-only export, and NVIDIA GPU selection on local machines or over SSH.">
<meta name="robots" content="index, follow, max-image-preview:large">
<meta name="theme-color" content="#010608">
<link rel="canonical" href="https://phantombugz.com/lived-lora-studio.html">
<meta property="og:site_name" content="PhantomBugz">
<meta property="og:title" content="liveD LoRA Studio — Developer Preview">
<meta property="og:description" content="Encrypted review and export for approved records, with NVIDIA GPU selection on local machines or over SSH.">
<meta property="og:type" content="website">
<meta property="og:url" content="https://phantombugz.com/lived-lora-studio.html">
<meta property="og:image" content="https://phantombugz.com/assets/exports/lived-social-preview.png">
<meta property="og:image:type" content="image/png">
<meta property="og:image:width" content="1280">
<meta property="og:image:height" content="640">
<meta name="twitter:card" content="summary_large_image">
<meta name="twitter:site" content="@phantombugz">
<meta name="twitter:title" content="liveD LoRA Studio — PhantomBugz">
<meta name="twitter:description" content="Encrypted review and approved-only export, with a separate NVIDIA GPU picker.">
<meta name="twitter:image" content="https://phantombugz.com/assets/exports/lived-social-preview.png">
<link rel="icon" href="./assets/exports/phantombugz-icon-32.png" sizes="32x32" type="image/png">
<link rel="icon" href="./assets/exports/phantombugz-icon-16.png" sizes="16x16" type="image/png">
<link rel="apple-touch-icon" href="./assets/exports/phantombugz-icon-180.png">
<link rel="preload" href="./assets/fonts/anton.woff2" as="font" type="font/woff2" crossorigin>
<link rel="preload" href="./assets/fonts/jetbrains-mono-400.woff2" as="font" type="font/woff2" crossorigin>
<link rel="stylesheet" href="./assets/fonts/fonts.css">
<link rel="stylesheet" href="./styles.css?v=a6f3b5e4">
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "SoftwareApplication",
"name": "liveD LoRA Studio",
"applicationCategory": "DeveloperApplication",
"operatingSystem": "Windows, Linux",
"license": "https://spdx.org/licenses/Apache-2.0.html",
"url": "https://phantombugz.com/lived-lora-studio.html",
"codeRepository": "https://github.com/PhantomBugz/lived-lora-studio",
"author": { "@type": "Organization", "name": "PhantomBugz", "url": "https://phantombugz.com/", "sameAs": ["https://x.com/phantombugz", "https://github.com/PhantomBugz"] }
}
</script>
</head>
<body class="zk-page lived-page">
<a class="skip-link" href="#lived-main">Skip to liveD LoRA Studio</a>
<header class="site-header" aria-label="Primary">
<a class="brand-lockup" href="./enter.html#arrival" aria-label="PhantomBugz home">
<img src="./assets/exports/phantombugz-emblem.svg" alt="" aria-hidden="true" width="34" height="34">
</a>
<nav class="header-nav" aria-label="Sections">
<a href="./enter.html">PhantomBugz</a>
<a href="./zombie-killer.html">Zombie Killer</a>
<a href="#quickstart">Quickstart</a>
<a class="social-link" href="https://github.com/PhantomBugz/lived-lora-studio" rel="noopener" target="_blank" aria-label="liveD LoRA Studio source on GitHub (opens in a new tab)"><svg class="social-icon" viewBox="0 0 16 16" width="18" height="18" aria-hidden="true" focusable="false"><path fill="currentColor" d="M8 0C3.58 0 0 3.58 0 8c0 3.54 2.29 6.53 5.47 7.59.4.07.55-.17.55-.38 0-.19-.01-.82-.01-1.49-2.01.37-2.53-.49-2.69-.94-.09-.23-.48-.94-.82-1.13-.28-.15-.68-.52-.01-.53.63-.01 1.08.58 1.23.82.72 1.21 1.87.87 2.33.66.07-.52.28-.87.51-1.07-1.78-.2-3.64-.89-3.64-3.95 0-.87.31-1.59.82-2.15-.08-.2-.36-1.02.08-2.12 0 0 .67-.21 2.2.82.64-.18 1.32-.27 2-.27s1.36.09 2 .27c1.53-1.04 2.2-.82 2.2-.82.44 1.1.16 1.92.08 2.12.51.56.82 1.27.82 2.15 0 3.07-1.87 3.75-3.65 3.95.29.25.54.73.54 1.48 0 1.07-.01 1.93-.01 2.2 0 .21.15.46.55.38A8.01 8.01 0 0 0 16 8c0-4.42-3.58-8-8-8Z"/></svg></a>
</nav>
</header>
<main id="lived-main" class="zk-shell" tabindex="-1">
<section class="zk-hero lived-hero" aria-labelledby="lived-title">
<img class="lived-hero-art" src="./assets/exports/lived-hero.png" width="1200" height="340" alt="PhantomBugz preview artwork showing encrypted review, approved-only export, and a separate local or SSH GPU picker">
<div class="lived-hero-copy">
<p class="eyebrow">Open source / developer preview</p>
<h1 id="lived-title">liveD LoRA Studio</h1>
<p class="zk-lead">Build training examples through explicit human review.</p>
<p class="zk-intro">Review records in an encrypted local vault and export approved records as encrypted JSONL. A separate picker locates local and remote NVIDIA GPUs by UUID.</p>
<p class="zk-status"><span>Source preview</span><span>Apache-2.0</span></p>
<div class="zk-actions">
<a class="zk-cta" href="https://github.com/PhantomBugz/lived-lora-studio" target="_blank" rel="noopener">Explore source on GitHub <span aria-hidden="true">→</span></a>
<a class="zk-text-link" href="#quickstart">Verify review and export</a>
</div>
</div>
</section>
<section id="quickstart" class="zk-capabilities lived-section" aria-labelledby="lived-quickstart-title">
<p class="eyebrow">Source quickstart</p>
<h2 id="lived-quickstart-title">Verify the export.</h2>
<p class="lived-section-lead">Run this command from the source checkout with Python 3.12 or newer. It creates a dedicated environment, installs the pinned cryptography dependency, and uses disposable synthetic data.</p>
<p class="lived-command"><code>python scripts/quickstart_preview.py</code></p>
<p class="lived-section-lead">The check confirms that only the approved record enters the encrypted JSONL export and that plaintext is absent from the staged ciphertext files.</p>
<figure class="lived-figure">
<img src="./assets/exports/lived-pipeline.png" width="1200" height="360" loading="lazy" alt="Preview workflow diagram: encrypted local vault, human review, approved-only encrypted export; the GPU target picker is separate and does not run training">
<figcaption>The encrypted review and export workflow. GPU selection runs separately.</figcaption>
</figure>
</section>
<section id="gpu-targets" class="zk-platforms lived-section" aria-labelledby="lived-gpu-title">
<p class="eyebrow">Exact target selection</p>
<h2 id="lived-gpu-title">Pick the machine. Pick the GPU.</h2>
<p class="lived-section-lead">The preview discovers NVIDIA devices on local Windows or Linux, or on a remote Linux machine reached through your existing SSH alias. It shows the UUID, device name, memory, compute capability, and driver, then requires an exact choice.</p>
<div class="zk-capability-grid lived-card-grid">
<article><p class="zk-card-number">01</p><h3>Local NVIDIA</h3><p>Inspect the cards visible to this machine and choose a specific UUID.</p></article>
<article><p class="zk-card-number">02</p><h3>SSH remote NVIDIA</h3><p>Use your configured SSH alias with strict host-key checking to inspect and select one remote UUID.</p></article>
</div>
<p class="lived-section-lead">The picker reads hardware inventory only. It does not send a dataset, install CUDA, or start model training. An observed GPU still needs a compatible runtime, model recipe, and execution smoke before training can be called ready.</p>
</section>
<section class="zk-release lived-section" aria-labelledby="lived-status-title">
<p class="eyebrow">Release boundary</p>
<h2 id="lived-status-title">Release status</h2>
<dl class="zk-platform-list lived-status-list">
<div><dt>Available in source</dt><dd>Encrypted review and approved-only export, plus NVIDIA GPU selection on local machines or over SSH.</dd></div>
<div><dt>Release limitations</dt><dd>GPU-specific CUDA training, the full desktop application, and platform installers remain unreleased.</dd></div>
</dl>
<p>This page describes the public source preview. It includes no collected corpus, private vault, model weights, or deployment credentials. Use only material you are authorized to collect and review.</p>
<p><a class="zk-text-link" href="https://github.com/PhantomBugz/lived-lora-studio/blob/main/docs/STATUS.md" target="_blank" rel="noopener">Read the current release status</a> or explore <a class="zk-text-link" href="./zombie-killer.html">Zombie Killer</a>, our separate process-safety project preview.</p>
</section>
</main>
<footer class="site-footer">
<a class="foot-mark" href="./enter.html" aria-label="PhantomBugz home">
<img src="./assets/exports/phantombugz-emblem.svg" alt="" aria-hidden="true" width="26" height="26">
<span>PhantomBugz</span>
</a>
<nav aria-label="Footer">
<a href="./enter.html">Home</a>
<a href="./zombie-killer.html">Zombie Killer</a>
<a href="https://github.com/PhantomBugz/lived-lora-studio" target="_blank" rel="noopener">liveD source</a>
</nav>
</footer>
</body>
</html>