ethanvert.github.io/rdmc.html at main · ethanvert/ethanvert.github.io · GitHub

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
<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Reverse Diffusion Monte Carlo</title>
    <link rel="stylesheet" href="./css/styles.css">
    <link rel="stylesheet" href="./css/project.css">
    <link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Open+Sans:wght@300;400;600;700&display=swap">
    <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0-beta3/css/all.min.css">
</head>
<body>

    <header>
        <div class="header-container">
            <h1><a href="index.html">Ethan Vertal</a></h1>
            <nav>
                <div class="menu-icon" onclick="toggleMenu()">
                    <i class="fas fa-bars"></i>
                </div>
                <ul id="nav-links" class="nav-links">
                    <li><a href="index.html#about-me">About Me</a></li>
                    <li><a href="resume.html">Resume</a></li>
                    <li><a href="index.html#projects">Projects</a></li>
                    <li><a href="index.html#contact">Contact</a></li>
                    <li><a href="./assets/Ethan_Vertal_Resume_Oct24.pdf" download>Download My Resume (PDF)</a></li>
                </ul>
            </nav>
        </div>
    </header>

    <div class="container">
        <section class="section" id="project">
            <h2>A New Method of Generating Images</h2>
            <div class="project-content">
                <img src="./assets/poster.png" alt="RDMC Poster" class="project-image">
                <p>
                    This project was our final assignment for Generative AI at CMU. We implemented a new way of looking at the reverse diffusion process from Huang et. al., 2024. To our knowledge, no one had implemented this before, so we decided it would be fun to try.
                </p>
                <p>
                    Implementing a new algorithm from just the paper comes with more challenges than we possibly could have accounted for. That being said, it was fun working with these two and I thought I'd make this work publicly available because I'm proud of what we did. Our results were below what we were hoping for, but Google Colab is not exactly friendly for 12+ hour coding sessions. Enjoy!
                </p>
                <div class="challenges-takeaways">
                    <div class="challenges">
                        <h4>Challenges</h4>
                        <ul>
                            <li>We needed to learn the underlying, complex math behind the algorithm.</li>
                            <li>Optimizing training around the constraints of Google Colab for extended coding sessions.</li>
                            <li>Collaborating and communicating effectively with multiple concurrent projects.</li>
                        </ul>
                    </div>
                    <div class="takeaways">
                        <h4>Takeaways</h4>
                        <ul>
                            <li>I deeply researched and understand the math behind multiple diffusion image generation algorithms.</li>
                            <li>In research, sometimes finding out what doesn't work is a step towards finding what does.</li>
                            <li>Working in a team you want to work for leads to much better results.</li>
                        </ul>
                    </div>
                </div>
                <h3>Technologies and Concepts Used:</h3>
                <ul>
                    <li>Python</li>
                    <li>PyTorch</li>
                    <li>Google Colab</li>
                    <li>UNet Neural Network</li>
                    <li>Reverse Diffusion Monte Carlo</li>
                    <li>Stochastic Differential Equations</li>
                    <li>Monte Carlo Methods</li>
                </ul>
                <h3>Learn More:</h3>
                <ul>
                    <a href="https://github.com/ethanvert/rdmc" target="_blank">Visit the Project GitHub</a>
                </ul>
            </div>
        </section>
    </div>

    <footer>
        <p>You can reach me at <a href="mailto:ethan.vertal@gmail.com">ethan.vertal@gmail.com</p>
    </footer>

    <script src="./js/script.js"></script>

</body>
</html>