-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathindex.html
More file actions
91 lines (78 loc) · 2.48 KB
/
index.html
File metadata and controls
91 lines (78 loc) · 2.48 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
<html lang="en">
<head>
<meta charset="utf-8"/>
<meta name="viewport" content="width=device-width,initial-scale=1"/>
<title>WebGPU MNIST</title>
<style>
* {
box-sizing: border-box;
}
body {
font-family: system-ui, -apple-system, sans-serif;
padding: 24px;
background: #f5f6f8;
color: #2d3748;
line-height: 1.6;
max-width: 800px;
margin: 0 auto;
}
h1 {
color: #1a202c;
font-weight: 600;
margin-bottom: 8px;
}
h3 {
color: #2d3748;
font-weight: 600;
margin: 0 0 12px 0;
}
p {
color: #4a5568;
margin: 8px 0;
}
a {
color: #2b6cb0;
text-decoration: none;
}
a:hover {
color: #1a4971;
text-decoration: underline;
}
.intro {
color: #4a5568;
margin-bottom: 24px;
}
.section {
background: #ffffff;
border: 1px solid #e2e8f0;
border-radius: 8px;
padding: 24px;
margin-bottom: 16px;
box-shadow: 0 1px 3px rgba(0, 0, 0, 0.05);
}
.section h3 {
border-bottom: 2px solid #e2e8f0;
padding-bottom: 12px;
margin-bottom: 12px;
}
.section .description {
color: #718096;
font-size: 0.9em;
}
</style>
</head>
<body>
<h1>WebGPU MNIST Demo</h1>
<p class="intro">A neural network trained entirely in-browser using WebGPU compute shaders.</p>
<div class="section">
<h3><a href="guess.html">Interactive Digit Recognizer</a></h3>
<p>Draw digits with your mouse and watch the pre-trained model predict in real-time. Visualizes data flowing through each layer of the neural network as you draw.</p>
<p class="description">Requires a modern browser with WebGPU support. Mobile would probably be unusable.</p>
</div>
<div class="section">
<h3><a href="train_and_test.html">Train and Test</a></h3>
<p>Train the MNIST model from scratch in your browser. Runs the full training loop on the 60,000 training images, then evaluates accuracy on the 10,000 test images.</p>
<p class="description">Runs 15 epochs with batch size 32. On M3 Max: ~8 minutes. With batch size 64: ~4 minutes. Can be trained in my iPhone 15 Pro.</p>
</div>
</body>
</html>