Kearm commited on
Commit
bfd1ece
·
verified ·
1 Parent(s): 3e61d9e

Upload benchmark.html with huggingface_hub

Browse files
Files changed (1) hide show
  1. benchmark.html +43 -0
benchmark.html ADDED
@@ -0,0 +1,43 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!doctype html>
2
+ <html lang="en">
3
+ <head>
4
+ <meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1">
5
+ <meta name="color-scheme" content="dark light">
6
+ <title>RESMP.DEV Benchmark Instrument</title>
7
+ <link rel="stylesheet" href="style.css"><link rel="stylesheet" href="benchmark.css"><link rel="stylesheet" href="benchmark-component.css">
8
+ <script type="module" src="benchmark-chart.js"></script>
9
+ </head>
10
+ <body class="benchmark-page">
11
+ <header class="bench-nav">
12
+ <a class="brand" href="index.html"><span class="mark">R/</span><strong>RESMP.DEV</strong></a>
13
+ <span>BENCHMARK INSTRUMENT / REV 01</span><a href="index.html">← ORG CARD</a>
14
+ </header>
15
+ <main class="bench-shell">
16
+ <div class="bench-intro">
17
+ <div><p class="eyebrow"><span></span> DATA-DRIVEN SMALL MULTIPLES</p><h1>Evaluation<br><em>instrument.</em></h1></div>
18
+ <div class="bench-note"><b>USAGE</b><p>Edit the JSON block in this page or call <code>chart.config = {...}</code>. Series order and colors stay fixed across every panel. Export production SVG from the chart.</p></div>
19
+ </div>
20
+ <resmp-benchmark-chart id="demo-chart">
21
+ <script type="application/json">
22
+ {"title":"MODEL SYSTEM EVALUATIONS","subtitle":"REFERENCE LAYOUT / DEMONSTRATION DATA","higherIsBetter":true,
23
+ "series":[{"id":"candidate","label":"CANDIDATE / MXFP4","color":"#6f36d9"},{"id":"control","label":"BF16 CONTROL","color":"#171719"},{"id":"native","label":"NATIVE RTN","color":"#707078"},{"id":"baseline","label":"BASELINE A","color":"#aaa7af"},{"id":"reference","label":"REFERENCE B","color":"#d5d2d8"}],
24
+ "groups":[
25
+ {"name":"STEM","panels":[{"name":"MATH EXACT","values":{"candidate":62.4,"control":58.9,"native":60.1,"baseline":54.2,"reference":57.8}},{"name":"CODE PASS@1","values":{"candidate":71.8,"control":67.3,"native":69.4,"baseline":61.5,"reference":64.9}},{"name":"SCIENCE QA","values":{"candidate":76.2,"control":73.7,"native":74.1,"baseline":69.8,"reference":72.5}}]},
26
+ {"name":"INSTRUCTION FOLLOWING","panels":[{"name":"FORMAT","values":{"candidate":84.6,"control":79.2,"native":81.4,"baseline":72.1,"reference":77.3}},{"name":"MULTI-TURN","values":{"candidate":68.5,"control":63.2,"native":65.9,"baseline":58.4,"reference":61.7}},{"name":"CONSTRAINTS","values":{"candidate":73.9,"control":70.6,"native":71.8,"baseline":66.2,"reference":68.5}}]},
27
+ {"name":"TOOL USE","panels":[{"name":"FUNCTION CALL","values":{"candidate":79.1,"control":70.8,"native":75.4,"baseline":64.2,"reference":69.1}},{"name":"SANDBOX","values":{"candidate":66.7,"control":61.1,"native":63.5,"baseline":56.9,"reference":60.2}},{"name":"RETRIEVAL","values":{"candidate":81.3,"control":76.5,"native":78.2,"baseline":71.8,"reference":75.6}}]},
28
+ {"name":"AGENTIC","panels":[{"name":"TASK COMPLETE","values":{"candidate":57.8,"control":49.6,"native":53.1,"baseline":45.4,"reference":50.2}},{"name":"PATCH ACCEPT","values":{"candidate":64.2,"control":55.8,"native":59.6,"baseline":51.7,"reference":56.1}},{"name":"BROWSER","values":{"candidate":47.9,"control":39.4,"native":43.3,"baseline":35.8,"reference":41.5}}]}
29
+ ]}
30
+ </script>
31
+ </resmp-benchmark-chart>
32
+ <section class="contract">
33
+ <span>INPUT CONTRACT</span>
34
+ <pre><code>{ title, subtitle, higherIsBetter,
35
+ series: [{ id, label, color }],
36
+ groups: [{ name, panels: [{
37
+ name, values: { [seriesId]: number }
38
+ }] }] }</code></pre>
39
+ <p>Plain Web Component. No framework, external font, chart library, or build step. Drop <code>benchmark-chart.js</code> and <code>benchmark-component.css</code> beside any static report.</p>
40
+ </section>
41
+ </main>
42
+ </body>
43
+ </html>