Spaces:

canada-quant
/

README

Running

App Files Files Community

pastapaul commited on 6 days ago

Commit

d646272

verified ·

1 Parent(s): 1d88966

Remove default static-template files so README.md renders as org card

Browse files

Files changed (2) hide show

index.html +0 -257
style.css +0 -28

index.html DELETED Viewed

@@ -1,257 +0,0 @@
-<!doctype html>
-<html lang="en">
-<head>
-    <meta charset="utf-8" />
-    <meta name="viewport" content="width=device-width, initial-scale=1" />
-    <title>Canada Quant Labs</title>
-    <style>
-        :root {
-            --red: #d52b1e;
-            --ink: #111;
-            --muted: #5a5a5a;
-            --rule: #e5e5e5;
-            --bg: #ffffff;
-            --code-bg: #f6f6f6;
-        }
-        * { box-sizing: border-box; }
-        html, body {
-            margin: 0;
-            padding: 0;
-            background: var(--bg);
-            color: var(--ink);
-            font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Inter, Roboto, "Helvetica Neue", Arial, sans-serif;
-            line-height: 1.55;
-            font-size: 15px;
-        }
-        .wrap {
-            max-width: 980px;
-            margin: 0 auto;
-            padding: 32px 24px 48px;
-        }
-        header {
-            border-bottom: 3px solid var(--red);
-            padding-bottom: 18px;
-            margin-bottom: 24px;
-        }
-        h1 {
-            font-size: 32px;
-            font-weight: 700;
-            margin: 0 0 6px;
-            letter-spacing: -0.01em;
-        }
-        h1 .leaf { color: var(--red); }
-        .tagline {
-            font-size: 16px;
-            color: var(--muted);
-            margin: 0;
-        }
-        h2 {
-            font-size: 20px;
-            font-weight: 600;
-            margin: 32px 0 12px;
-            letter-spacing: -0.005em;
-        }
-        h3 {
-            font-size: 15px;
-            font-weight: 600;
-            margin: 22px 0 8px;
-            color: var(--muted);
-            text-transform: uppercase;
-            letter-spacing: 0.04em;
-        }
-        p { margin: 0 0 12px; }
-        ul {
-            margin: 0 0 14px;
-            padding-left: 20px;
-        }
-        li { margin-bottom: 4px; }
-        a {
-            color: var(--red);
-            text-decoration: none;
-            border-bottom: 1px solid transparent;
-            transition: border-color 0.15s;
-        }
-        a:hover { border-bottom-color: var(--red); }
-        code {
-            background: var(--code-bg);
-            padding: 1px 5px;
-            border-radius: 3px;
-            font-family: ui-monospace, "SF Mono", Menlo, Consolas, monospace;
-            font-size: 0.9em;
-        }
-        .pillars {
-            display: grid;
-            grid-template-columns: repeat(auto-fit, minmax(220px, 1fr));
-            gap: 18px;
-            margin: 18px 0 8px;
-        }
-        .pillar {
-            border-left: 3px solid var(--red);
-            padding: 4px 0 4px 14px;
-        }
-        .pillar-title {
-            font-weight: 600;
-            font-size: 13px;
-            color: var(--muted);
-            text-transform: uppercase;
-            letter-spacing: 0.04em;
-            margin-bottom: 6px;
-        }
-        .pillar-body { font-size: 14px; }
-        .table-scroll {
-            overflow-x: auto;
-            margin: 8px 0 16px;
-            border: 1px solid var(--rule);
-            border-radius: 6px;
-        }
-        table {
-            width: 100%;
-            border-collapse: collapse;
-            font-size: 13.5px;
-        }
-        th, td {
-            text-align: left;
-            padding: 10px 12px;
-            border-bottom: 1px solid var(--rule);
-            vertical-align: top;
-        }
-        th {
-            background: #fafafa;
-            font-weight: 600;
-            white-space: nowrap;
-        }
-        tr:last-child td { border-bottom: none; }
-        td code { font-size: 12.5px; }
-        .meta-line {
-            font-size: 13.5px;
-            color: var(--muted);
-            margin: 14px 0;
-        }
-        .meta-line strong { color: var(--ink); }
-        .contact {
-            margin-top: 28px;
-            padding-top: 18px;
-            border-top: 1px solid var(--rule);
-            font-size: 14px;
-            color: var(--muted);
-        }
-        .contact a { color: var(--ink); }
-        .contact a:hover { color: var(--red); border-bottom-color: var(--red); }
-        .repos-list {
-            font-size: 13.5px;
-            list-style: none;
-            padding-left: 0;
-            margin-top: 4px;
-        }
-        .repos-list li { margin-bottom: 3px; }
-    </style>
-</head>
-<body>
-<div class="wrap">
-    <header>
-        <h1><span class="leaf">🍁</span> Canada Quant Labs</h1>
-        <p class="tagline">Canada's open-weight model lab. Sovereign AI on Canadian Blackwell silicon for regulated industries.</p>
-    </header>
-    <div class="pillars">
-        <div class="pillar">
-            <div class="pillar-title">What we do</div>
-            <div class="pillar-body">Post-training on open base models (SFT, DPO, GRPO, RLAIF) · Production quantization recipes (W4A16, NVFP4, MXFP4) · Audited, air-gapped deployment with eval evidence and MRM docs.</div>
-        </div>
-        <div class="pillar">
-            <div class="pillar-title">Where we work</div>
-            <div class="pillar-body">Legal · Medical · Defence · Finance. Headquarters in Victoria, BC. Compute on NVIDIA DGX B300 at Equinix Vancouver.</div>
-        </div>
-        <div class="pillar">
-            <div class="pillar-title">Upstream</div>
-            <div class="pillar-body">Contributors to vLLM, llm-compressor, compressed-tensors. PR <a href="https://github.com/vllm-project/vllm/pull/42209">#42209</a> (NVFP4 MoE for DSV4) merged 2026-05-22 with our validation.</div>
-        </div>
-    </div>
-    <h2>Open releases — DeepSeek-V4 quantization family</h2>
-    <p>Four artifacts in the same lineage. One base model in two sizes (V4-Flash, V4-Pro); two routed-expert formats (W4A16, NVFP4); Multi-Token Prediction (MTP) draft head retained on three of four. Attention is FP8 block 128×128 across all four.</p>
-    <div class="table-scroll">
-    <table>
-        <thead>
-            <tr>
-                <th>Model</th>
-                <th>Base</th>
-                <th>Routed experts</th>
-                <th>MTP</th>
-                <th>On-disk</th>
-                <th>Min hardware (TP=2)</th>
-                <th>When to pick</th>
-            </tr>
-        </thead>
-        <tbody>
-            <tr>
-                <td><a href="https://huggingface.co/canada-quant/DeepSeek-V4-Flash-W4A16-FP8">V4-Flash-W4A16-FP8</a></td>
-                <td>V4-Flash</td>
-                <td>W4A16 INT4 g=128</td>
-                <td>no</td>
-                <td>~143 GB</td>
-                <td>H200 / DGX Spark / RTX PRO 6000</td>
-                <td>maximum compatibility, no MTP needed</td>
-            </tr>
-            <tr>
-                <td><a href="https://huggingface.co/canada-quant/DeepSeek-V4-Flash-W4A16-FP8-MTP">V4-Flash-W4A16-FP8-MTP</a></td>
-                <td>V4-Flash</td>
-                <td>W4A16 INT4 g=128</td>
-                <td>yes (BF16)</td>
-                <td>159 GB</td>
-                <td>H200 / RTX PRO 6000</td>
-                <td>best $/token interactive on V4-Flash</td>
-            </tr>
-            <tr>
-                <td><a href="https://huggingface.co/canada-quant/DeepSeek-V4-Flash-NVFP4-FP8-MTP">V4-Flash-NVFP4-FP8-MTP</a></td>
-                <td>V4-Flash</td>
-                <td>NVFP4 g=16</td>
-                <td>yes (BF16)</td>
-                <td>172 GB</td>
-                <td>RTX PRO 6000 / B300</td>
-                <td>best Blackwell-native interactive on V4-Flash</td>
-            </tr>
-            <tr>
-                <td><a href="https://huggingface.co/canada-quant/DeepSeek-V4-Pro-NVFP4-FP8-MTP">V4-Pro-NVFP4-FP8-MTP</a></td>
-                <td>V4-Pro</td>
-                <td>NVFP4 g=16</td>
-                <td>yes (byte-identical)</td>
-                <td>913 GiB</td>
-                <td>8× B300 (TP=8 + EP)</td>
-                <td>only choice for V4-Pro; <strong>+25–37% throughput vs upstream MXFP4</strong></td>
-            </tr>
-        </tbody>
-    </table>
-    </div>
-    <p class="meta-line">Upstream reference recipes: <a href="https://huggingface.co/RedHatAI/DeepSeek-V4-Flash-NVFP4-FP8"><code>RedHatAI/DeepSeek-V4-Flash-NVFP4-FP8</code></a> (Flash NVFP4 topology) and <a href="https://huggingface.co/nvidia/DeepSeek-V3.2-NVFP4"><code>nvidia/DeepSeek-V3.2-NVFP4</code></a> (Pro NVFP4, MTP-exclusion topology).</p>
-    <h3>Hardware shorthand</h3>
-    <ul>
-        <li><strong>H200</strong> — 8× NVIDIA H200 SXM5 (Hopper SM 9.0a, 141 GB HBM3e/GPU)</li>
-        <li><strong>DGX Spark</strong> — 2× NVIDIA DGX Spark (GB10, Blackwell SM 12.1a)</li>
-        <li><strong>RTX PRO 6000</strong> — NVIDIA RTX PRO 6000 Blackwell Server Edition (SM 12.0, sm_120, 96 GB HBM)</li>
-        <li><strong>B300</strong> — NVIDIA B300 SXM6 AC (Blackwell SM 10.3, sm_103a, 288 GB HBM3e/GPU)</li>
-    </ul>
-    <h3>Reproduction repos</h3>
-    <p>Every artifact has a public reproduction repo with calibration scripts, vLLM patches, bench harnesses, and findings docs:</p>
-    <ul class="repos-list">
-        <li>↳ <a href="https://github.com/canada-quant/dsv4-flash-w4a16-fp8"><code>canada-quant/dsv4-flash-w4a16-fp8</code></a></li>
-        <li>↳ <a href="https://github.com/canada-quant/dsv4-flash-w4a16-fp8-mtp"><code>canada-quant/dsv4-flash-w4a16-fp8-mtp</code></a></li>
-        <li>↳ <a href="https://github.com/canada-quant/dsv4-flash-nvfp4-fp8-mtp"><code>canada-quant/dsv4-flash-nvfp4-fp8-mtp</code></a></li>
-        <li>↳ <a href="https://github.com/canada-quant/dsv4-pro-nvfp4-fp8-mtp"><code>canada-quant/dsv4-pro-nvfp4-fp8-mtp</code></a></li>
-    </ul>
-    <div class="contact">
-        Partnerships · <a href="mailto:partnerships@cql.ca">partnerships@cql.ca</a> &nbsp;&middot;&nbsp;
-        Press · <a href="mailto:press@cql.ca">press@cql.ca</a> &nbsp;&middot;&nbsp;
-        Web · <a href="https://cql.ca">cql.ca</a>
-    </div>
-</div>
-</body>
-</html>

style.css DELETED Viewed

@@ -1,28 +0,0 @@
-body {
-	padding: 2rem;
-	font-family: -apple-system, BlinkMacSystemFont, "Arial", sans-serif;
-}
-h1 {
-	font-size: 16px;
-	margin-top: 0;
-}
-p {
-	color: rgb(107, 114, 128);
-	font-size: 15px;
-	margin-bottom: 10px;
-	margin-top: 5px;
-}
-.card {
-	max-width: 620px;
-	margin: 0 auto;
-	padding: 16px;
-	border: 1px solid lightgray;
-	border-radius: 16px;
-}
-.card p:last-child {
-	margin-bottom: 0;
-}