Spaces:
Running
Running
Test new design template.
#1
by
metacritical
- opened
- index.html +106 -175
index.html
CHANGED
|
@@ -1,188 +1,119 @@
|
|
| 1 |
<!DOCTYPE html>
|
| 2 |
-
<html>
|
| 3 |
<head>
|
| 4 |
-
<meta charset="
|
| 5 |
-
<meta name="
|
| 6 |
-
<
|
| 7 |
-
<
|
| 8 |
-
<
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
</head>
|
| 25 |
<body>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
<
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
<div class="is-size-5 publication-authors">
|
| 34 |
-
A collection of groundbreaking research papers in AI and language models
|
| 35 |
-
</div>
|
| 36 |
-
</div>
|
| 37 |
-
</div>
|
| 38 |
</div>
|
| 39 |
-
|
| 40 |
-
</
|
| 41 |
-
|
| 42 |
-
<
|
| 43 |
-
|
| 44 |
-
<!-- Abstract. -->
|
| 45 |
-
<div class="columns is-centered has-text-centered">
|
| 46 |
-
<div class="column is-four-fifths">
|
| 47 |
-
<h2 class="title is-3">Overview</h2>
|
| 48 |
-
<div class="content has-text-justified">
|
| 49 |
-
<p>
|
| 50 |
-
DeepSeek has released a series of significant papers detailing advancements in large language models (LLMs).
|
| 51 |
-
Each paper represents a step forward in making AI more capable, efficient, and accessible.
|
| 52 |
-
</p>
|
| 53 |
-
</div>
|
| 54 |
-
</div>
|
| 55 |
</div>
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
<h2 class="title is-3">Research Papers</h2>
|
| 62 |
-
|
| 63 |
-
<!-- Paper 1 -->
|
| 64 |
-
<div class="publication-block">
|
| 65 |
-
<div class="publication-header">
|
| 66 |
-
<h3 class="title is-4">DeepSeekLLM: Scaling Open-Source Language Models with Longer-termism</h3>
|
| 67 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 68 |
-
<div class="is-size-5 publication-authors">
|
| 69 |
-
Released: November 29, 2023
|
| 70 |
-
</div>
|
| 71 |
-
</div>
|
| 72 |
-
<div class="content has-text-justified">
|
| 73 |
-
<p>This foundational paper explores scaling laws and the trade-offs between data and model size,
|
| 74 |
-
establishing the groundwork for subsequent models.</p>
|
| 75 |
-
</div>
|
| 76 |
-
</div>
|
| 77 |
-
|
| 78 |
-
<!-- Paper 2 -->
|
| 79 |
-
<div class="publication-block">
|
| 80 |
-
<div class="publication-header">
|
| 81 |
-
<h3 class="title is-4">DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model</h3>
|
| 82 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 83 |
-
<div class="is-size-5 publication-authors">
|
| 84 |
-
Released: May 2024
|
| 85 |
-
</div>
|
| 86 |
-
</div>
|
| 87 |
-
<div class="content has-text-justified">
|
| 88 |
-
<p>Introduces a Mixture-of-Experts (MoE) architecture, enhancing performance while reducing
|
| 89 |
-
training costs by 42%.</p>
|
| 90 |
-
</div>
|
| 91 |
-
</div>
|
| 92 |
-
|
| 93 |
-
<!-- Additional papers following same structure -->
|
| 94 |
-
<div class="publication-block">
|
| 95 |
-
<div class="publication-header">
|
| 96 |
-
<h3 class="title is-4">DeepSeek-V3 Technical Report</h3>
|
| 97 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 98 |
-
<div class="is-size-5 publication-authors">
|
| 99 |
-
Released: December 2024
|
| 100 |
-
</div>
|
| 101 |
-
</div>
|
| 102 |
-
<div class="content has-text-justified">
|
| 103 |
-
<p>Discusses the scaling of sparse MoE networks to 671 billion parameters.</p>
|
| 104 |
-
</div>
|
| 105 |
-
</div>
|
| 106 |
-
|
| 107 |
-
<div class="publication-block">
|
| 108 |
-
<div class="publication-header">
|
| 109 |
-
<h3 class="title is-4">DeepSeek-R1: Incentivizing Reasoning Capability in LLMs</h3>
|
| 110 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 111 |
-
<div class="is-size-5 publication-authors">
|
| 112 |
-
Released: January 20, 2025
|
| 113 |
-
</div>
|
| 114 |
-
</div>
|
| 115 |
-
<div class="content has-text-justified">
|
| 116 |
-
<p>Enhances reasoning capabilities through large-scale reinforcement learning.</p>
|
| 117 |
-
</div>
|
| 118 |
-
</div>
|
| 119 |
-
|
| 120 |
-
<div class="publication-block">
|
| 121 |
-
<div class="publication-header">
|
| 122 |
-
<h3 class="title is-4">DeepSeekMath: Pushing the Limits of Mathematical Reasoning</h3>
|
| 123 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 124 |
-
<div class="is-size-5 publication-authors">
|
| 125 |
-
Released: April 2024
|
| 126 |
-
</div>
|
| 127 |
-
</div>
|
| 128 |
-
<div class="content has-text-justified">
|
| 129 |
-
<p>Presents methods to improve mathematical reasoning in LLMs.</p>
|
| 130 |
-
</div>
|
| 131 |
-
</div>
|
| 132 |
-
|
| 133 |
-
<div class="publication-block">
|
| 134 |
-
<div class="publication-header">
|
| 135 |
-
<h3 class="title is-4">DeepSeek-Prover: Advancing Theorem Proving in LLMs</h3>
|
| 136 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 137 |
-
</div>
|
| 138 |
-
<div class="content has-text-justified">
|
| 139 |
-
<p>Focuses on enhancing theorem proving capabilities using synthetic data for training.</p>
|
| 140 |
-
</div>
|
| 141 |
-
</div>
|
| 142 |
-
|
| 143 |
-
<div class="publication-block">
|
| 144 |
-
<div class="publication-header">
|
| 145 |
-
<h3 class="title is-4">DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models</h3>
|
| 146 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 147 |
-
</div>
|
| 148 |
-
<div class="content has-text-justified">
|
| 149 |
-
<p>Details advancements in code-related tasks with emphasis on open-source methodologies.</p>
|
| 150 |
-
</div>
|
| 151 |
-
</div>
|
| 152 |
-
|
| 153 |
-
<div class="publication-block">
|
| 154 |
-
<div class="publication-header">
|
| 155 |
-
<h3 class="title is-4">DeepSeekMoE: Advancing Mixture-of-Experts Architecture</h3>
|
| 156 |
-
<span class="tag is-primary is-medium">Deep Dive Coming Soon</span>
|
| 157 |
-
</div>
|
| 158 |
-
<div class="content has-text-justified">
|
| 159 |
-
<p>Discusses the integration and benefits of the Mixture-of-Experts approach.</p>
|
| 160 |
-
</div>
|
| 161 |
-
</div>
|
| 162 |
-
</div>
|
| 163 |
</div>
|
| 164 |
-
|
| 165 |
-
</
|
| 166 |
-
|
| 167 |
-
<
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
</
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 173 |
</div>
|
| 174 |
-
<div class="
|
| 175 |
-
<
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
</div>
|
| 184 |
</div>
|
| 185 |
-
|
| 186 |
-
|
|
|
|
| 187 |
</body>
|
| 188 |
</html>
|
|
|
|
| 1 |
<!DOCTYPE html>
|
| 2 |
+
<html lang="en">
|
| 3 |
<head>
|
| 4 |
+
<meta charset="UTF-8">
|
| 5 |
+
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
| 6 |
+
<title>DeepSeek Papers</title>
|
| 7 |
+
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0-beta3/css/all.min.css">
|
| 8 |
+
<style>
|
| 9 |
+
body {
|
| 10 |
+
font-family: 'Arial', sans-serif;
|
| 11 |
+
margin: 0;
|
| 12 |
+
padding: 0;
|
| 13 |
+
line-height: 1.6;
|
| 14 |
+
color: #333;
|
| 15 |
+
background-color: #f9f9f9;
|
| 16 |
+
}
|
| 17 |
+
header {
|
| 18 |
+
background: #4CAF50;
|
| 19 |
+
color: white;
|
| 20 |
+
padding: 20px 0;
|
| 21 |
+
text-align: center;
|
| 22 |
+
}
|
| 23 |
+
h1 {
|
| 24 |
+
margin: 0;
|
| 25 |
+
font-size: 2.5em;
|
| 26 |
+
}
|
| 27 |
+
.container {
|
| 28 |
+
max-width: 800px;
|
| 29 |
+
margin: 20px auto;
|
| 30 |
+
padding: 20px;
|
| 31 |
+
background: white;
|
| 32 |
+
border-radius: 8px;
|
| 33 |
+
box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
|
| 34 |
+
}
|
| 35 |
+
.paper {
|
| 36 |
+
margin-bottom: 20px;
|
| 37 |
+
}
|
| 38 |
+
.paper a {
|
| 39 |
+
text-decoration: none;
|
| 40 |
+
color: #4CAF50;
|
| 41 |
+
font-weight: bold;
|
| 42 |
+
}
|
| 43 |
+
.paper a:hover {
|
| 44 |
+
text-decoration: underline;
|
| 45 |
+
}
|
| 46 |
+
.coming-soon {
|
| 47 |
+
color: #e74c3c;
|
| 48 |
+
font-size: 0.9em;
|
| 49 |
+
margin-left: 10px;
|
| 50 |
+
}
|
| 51 |
+
footer {
|
| 52 |
+
text-align: center;
|
| 53 |
+
padding: 10px 0;
|
| 54 |
+
background: #4CAF50;
|
| 55 |
+
color: white;
|
| 56 |
+
margin-top: 20px;
|
| 57 |
+
}
|
| 58 |
+
</style>
|
| 59 |
</head>
|
| 60 |
<body>
|
| 61 |
+
<header>
|
| 62 |
+
<h1>DeepSeek Papers</h1>
|
| 63 |
+
</header>
|
| 64 |
+
<div class="container">
|
| 65 |
+
<h2>DeepSeek Research Contributions</h2>
|
| 66 |
+
<p>Below is a list of significant papers by DeepSeek detailing advancements in large language models (LLMs). Each paper includes a brief description and highlights upcoming deep dives.</p>
|
| 67 |
|
| 68 |
+
<!-- Paper List -->
|
| 69 |
+
<div class="paper">
|
| 70 |
+
<a href="#">DeepSeekLLM: Scaling Open-Source Language Models with Longer-termism</a>
|
| 71 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 72 |
+
<p><strong>Release Date:</strong> November 29, 2023<br>
|
| 73 |
+
This foundational paper explores scaling laws and the trade-offs between data and model size, establishing the groundwork for subsequent models.</p>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
</div>
|
| 75 |
+
<div class="paper">
|
| 76 |
+
<a href="#">DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model</a>
|
| 77 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 78 |
+
<p><strong>Release Date:</strong> May 2024<br>
|
| 79 |
+
This paper introduces a Mixture-of-Experts (MoE) architecture, enhancing performance while reducing training costs by 42%.</p>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
</div>
|
| 81 |
+
<div class="paper">
|
| 82 |
+
<a href="#">DeepSeek-V3 Technical Report</a>
|
| 83 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 84 |
+
<p><strong>Release Date:</strong> December 2024<br>
|
| 85 |
+
This report discusses the scaling of sparse MoE networks to 671 billion parameters, utilizing mixed precision training and HPC co-design strategies.</p>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 86 |
</div>
|
| 87 |
+
<div class="paper">
|
| 88 |
+
<a href="#">DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning</a>
|
| 89 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 90 |
+
<p><strong>Release Date:</strong> January 20, 2025<br>
|
| 91 |
+
The R1 model enhances reasoning capabilities through large-scale reinforcement learning, competing directly with leading models like OpenAI's o1.</p>
|
| 92 |
+
</div>
|
| 93 |
+
<div class="paper">
|
| 94 |
+
<a href="#">DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models</a>
|
| 95 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 96 |
+
<p><strong>Release Date:</strong> April 2024<br>
|
| 97 |
+
This paper presents methods to improve mathematical reasoning in LLMs, introducing the Group Relative Policy Optimization (GRPO) algorithm.</p>
|
| 98 |
+
</div>
|
| 99 |
+
<div class="paper">
|
| 100 |
+
<a href="#">DeepSeek-Prover: Advancing Theorem Proving in LLMs through Large-Scale Synthetic Data</a>
|
| 101 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 102 |
+
<p>Focuses on enhancing theorem proving capabilities in language models using synthetic data for training.</p>
|
| 103 |
</div>
|
| 104 |
+
<div class="paper">
|
| 105 |
+
<a href="#">DeepSeek-Coder-V2: Breaking the Barrier of Closed-Source Models in Code Intelligence</a>
|
| 106 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 107 |
+
<p>This paper details advancements in code-related tasks with an emphasis on open-source methodologies, improving upon earlier coding models.</p>
|
| 108 |
+
</div>
|
| 109 |
+
<div class="paper">
|
| 110 |
+
<a href="#">DeepSeekMoE</a>
|
| 111 |
+
<span class="coming-soon">[Deep Dive Coming Soon]</span>
|
| 112 |
+
<p>Discusses the integration and benefits of the Mixture-of-Experts approach within the DeepSeek framework.</p>
|
| 113 |
</div>
|
| 114 |
</div>
|
| 115 |
+
<footer>
|
| 116 |
+
© 2025 DeepSeek Research. All rights reserved.
|
| 117 |
+
</footer>
|
| 118 |
</body>
|
| 119 |
</html>
|