Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
|
@@ -9,6 +9,177 @@ pinned: false
|
|
| 9 |
|
| 10 |
<img src="https://huggingface.co/datasets/nanochat-students/images/resolve/main/students.png" alt="nanochat students banner" style="width: 100%; height: 500px; object-fit: cover; object-position: center;">
|
| 11 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
# nanochat students
|
| 13 |
|
| 14 |
Welcome to the **nanochat students** organization\! This is a community organization for students following Andrej Karpathy's [nanochat course](https://github.com/karpathy/nanochat). We are learning to build a full-stack LLM implementation from tokenization to web serving, all for under $100.
|
|
|
|
| 9 |
|
| 10 |
<img src="https://huggingface.co/datasets/nanochat-students/images/resolve/main/students.png" alt="nanochat students banner" style="width: 100%; height: 500px; object-fit: cover; object-position: center;">
|
| 11 |
|
| 12 |
+
<!DOCTYPE html>
|
| 13 |
+
<html lang="en">
|
| 14 |
+
<head>
|
| 15 |
+
<meta charset="UTF-8">
|
| 16 |
+
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
| 17 |
+
<title>Day 1 of Nano Chat</title>
|
| 18 |
+
<style>
|
| 19 |
+
* {
|
| 20 |
+
margin: 0;
|
| 21 |
+
padding: 0;
|
| 22 |
+
box-sizing: border-box;
|
| 23 |
+
}
|
| 24 |
+
|
| 25 |
+
body {
|
| 26 |
+
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, Oxygen, Ubuntu, Cantarell, sans-serif;
|
| 27 |
+
background: lightgrey;
|
| 28 |
+
min-height: 100vh;
|
| 29 |
+
display: flex;
|
| 30 |
+
align-items: center;
|
| 31 |
+
justify-content: center;
|
| 32 |
+
padding: 20px;
|
| 33 |
+
}
|
| 34 |
+
|
| 35 |
+
.container {
|
| 36 |
+
background: white;
|
| 37 |
+
border-radius: 16px;
|
| 38 |
+
box-shadow: 0 20px 60px rgba(0, 0, 0, 0.3);
|
| 39 |
+
max-width: 600px;
|
| 40 |
+
width: 100%;
|
| 41 |
+
overflow: hidden;
|
| 42 |
+
}
|
| 43 |
+
|
| 44 |
+
.header {
|
| 45 |
+
background: grey;
|
| 46 |
+
color: white;
|
| 47 |
+
padding: 30px;
|
| 48 |
+
text-align: center;
|
| 49 |
+
}
|
| 50 |
+
|
| 51 |
+
.header h1 {
|
| 52 |
+
font-size: 28px;
|
| 53 |
+
font-weight: 700;
|
| 54 |
+
margin-bottom: 8px;
|
| 55 |
+
}
|
| 56 |
+
|
| 57 |
+
.header p {
|
| 58 |
+
font-size: 14px;
|
| 59 |
+
opacity: 0.95;
|
| 60 |
+
}
|
| 61 |
+
|
| 62 |
+
.content {
|
| 63 |
+
padding: 30px;
|
| 64 |
+
}
|
| 65 |
+
|
| 66 |
+
.step {
|
| 67 |
+
margin-bottom: 24px;
|
| 68 |
+
padding: 20px;
|
| 69 |
+
background: #f8f9fa;
|
| 70 |
+
border-radius: 12px;
|
| 71 |
+
border-left: 4px solid #667eea;
|
| 72 |
+
transition: transform 0.2s, box-shadow 0.2s;
|
| 73 |
+
}
|
| 74 |
+
|
| 75 |
+
.step:hover {
|
| 76 |
+
transform: translateX(4px);
|
| 77 |
+
box-shadow: 0 4px 12px rgba(102, 126, 234, 0.15);
|
| 78 |
+
}
|
| 79 |
+
|
| 80 |
+
.step:last-child {
|
| 81 |
+
margin-bottom: 0;
|
| 82 |
+
}
|
| 83 |
+
|
| 84 |
+
.step-number {
|
| 85 |
+
display: inline-block;
|
| 86 |
+
background: #667eea;
|
| 87 |
+
color: white;
|
| 88 |
+
width: 28px;
|
| 89 |
+
height: 28px;
|
| 90 |
+
border-radius: 50%;
|
| 91 |
+
text-align: center;
|
| 92 |
+
line-height: 28px;
|
| 93 |
+
font-weight: 700;
|
| 94 |
+
font-size: 14px;
|
| 95 |
+
margin-right: 12px;
|
| 96 |
+
}
|
| 97 |
+
|
| 98 |
+
.step-title {
|
| 99 |
+
font-size: 18px;
|
| 100 |
+
font-weight: 600;
|
| 101 |
+
color: #2d3748;
|
| 102 |
+
margin-bottom: 8px;
|
| 103 |
+
}
|
| 104 |
+
|
| 105 |
+
.step-description {
|
| 106 |
+
color: #4a5568;
|
| 107 |
+
line-height: 1.6;
|
| 108 |
+
font-size: 15px;
|
| 109 |
+
}
|
| 110 |
+
|
| 111 |
+
.step-link {
|
| 112 |
+
display: inline-block;
|
| 113 |
+
margin-top: 12px;
|
| 114 |
+
color: #667eea;
|
| 115 |
+
text-decoration: none;
|
| 116 |
+
font-weight: 500;
|
| 117 |
+
font-size: 14px;
|
| 118 |
+
transition: color 0.2s;
|
| 119 |
+
}
|
| 120 |
+
|
| 121 |
+
.step-link:hover {
|
| 122 |
+
color: #764ba2;
|
| 123 |
+
text-decoration: underline;
|
| 124 |
+
}
|
| 125 |
+
|
| 126 |
+
.emoji {
|
| 127 |
+
font-size: 22px;
|
| 128 |
+
margin-right: 8px;
|
| 129 |
+
}
|
| 130 |
+
</style>
|
| 131 |
+
</head>
|
| 132 |
+
<body>
|
| 133 |
+
<div class="container">
|
| 134 |
+
<div class="header">
|
| 135 |
+
<h1>Day 1 of Nano Chat</h1>
|
| 136 |
+
<p>Getting started with language model training</p>
|
| 137 |
+
</div>
|
| 138 |
+
|
| 139 |
+
<div class="content">
|
| 140 |
+
<div class="step">
|
| 141 |
+
<div class="step-title">
|
| 142 |
+
<span class="step-number">1</span>
|
| 143 |
+
<span class="emoji">🛠️</span>Environment Setup
|
| 144 |
+
</div>
|
| 145 |
+
<div class="step-description">
|
| 146 |
+
Set up your Python environment using uv, create a virtual environment, and install all necessary dependencies for the nanochat project.
|
| 147 |
+
</div>
|
| 148 |
+
<a href="https://huggingface.co/spaces/nanochat-students/README/discussions/6" class="step-link" target="_blank">
|
| 149 |
+
View setup instructions →
|
| 150 |
+
</a>
|
| 151 |
+
</div>
|
| 152 |
+
|
| 153 |
+
<div class="step">
|
| 154 |
+
<div class="step-title">
|
| 155 |
+
<span class="step-number">2</span>
|
| 156 |
+
<span class="emoji">🔪</span>Tokenizer Training
|
| 157 |
+
</div>
|
| 158 |
+
<div class="step-description">
|
| 159 |
+
Train a custom BPE tokenizer using Rust bindings on 2 billion characters of data, achieving competitive compression ratios compared to GPT-4's tokenizer.
|
| 160 |
+
</div>
|
| 161 |
+
<a href="https://huggingface.co/spaces/nanochat-students/README/discussions/3" class="step-link" target="_blank">
|
| 162 |
+
View tokenizer guide →
|
| 163 |
+
</a>
|
| 164 |
+
</div>
|
| 165 |
+
|
| 166 |
+
<div class="step">
|
| 167 |
+
<div class="step-title">
|
| 168 |
+
<span class="step-number">3</span>
|
| 169 |
+
<span class="emoji">🥦</span>Pre-training
|
| 170 |
+
</div>
|
| 171 |
+
<div class="step-description">
|
| 172 |
+
Download a larger dataset and run distributed training across 8 GPUs using torchrun, with metrics tracked in a shared trackio space.
|
| 173 |
+
</div>
|
| 174 |
+
<a href="https://huggingface.co/spaces/nanochat-students/README/discussions/2" class="step-link" target="_blank">
|
| 175 |
+
View pre-training steps →
|
| 176 |
+
</a>
|
| 177 |
+
</div>
|
| 178 |
+
</div>
|
| 179 |
+
</div>
|
| 180 |
+
</body>
|
| 181 |
+
</html>
|
| 182 |
+
|
| 183 |
# nanochat students
|
| 184 |
|
| 185 |
Welcome to the **nanochat students** organization\! This is a community organization for students following Andrej Karpathy's [nanochat course](https://github.com/karpathy/nanochat). We are learning to build a full-stack LLM implementation from tokenization to web serving, all for under $100.
|