Deploy Scaling RL for LLMs deck (AMD AI Dev Day)
Browse files- .gitattributes +5 -0
- README.md +18 -4
- assets/adithyask-CfgQu2fr.jpeg +0 -0
- assets/blog-hero-A-OkyEfb.png +3 -0
- assets/cheating-agents-Bpt4f8GA.jpeg +0 -0
- assets/cognitivelab-BpHudwdd.png +0 -0
- assets/index-BaIqJAjN.js +0 -0
- assets/index-DlQWIdL0.css +1 -0
- assets/openai-ss1-DOT2K7xh.png +3 -0
- assets/openai-ss2-crop-BqCLkRBg.png +0 -0
- assets/openai-ss3-crop-KZZxFTHs.png +3 -0
- assets/qwen-scaling-C1rB1MJm.png +3 -0
- assets/reward-hack-curves-Dsoa5-YV.png +3 -0
- index.html +41 -17
- og.png +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
assets/blog-hero-A-OkyEfb.png filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
assets/openai-ss1-DOT2K7xh.png filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
assets/openai-ss3-crop-KZZxFTHs.png filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
assets/qwen-scaling-C1rB1MJm.png filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
assets/reward-hack-curves-Dsoa5-YV.png filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
|
@@ -1,10 +1,24 @@
|
|
| 1 |
---
|
| 2 |
-
title: Scaling
|
| 3 |
-
emoji:
|
| 4 |
colorFrom: green
|
| 5 |
-
colorTo:
|
| 6 |
sdk: static
|
|
|
|
| 7 |
pinned: false
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Scaling RL for LLMs — Slides
|
| 3 |
+
emoji: 📈
|
| 4 |
colorFrom: green
|
| 5 |
+
colorTo: purple
|
| 6 |
sdk: static
|
| 7 |
+
app_file: index.html
|
| 8 |
pinned: false
|
| 9 |
+
license: mit
|
| 10 |
+
short_description: Scaling RL for LLMs — AMD AI Dev Day slides
|
| 11 |
---
|
| 12 |
|
| 13 |
+
# Scaling RL for LLMs — RL Environments and RL Training
|
| 14 |
+
|
| 15 |
+
Talk slides by [Adithya S Kolavi](https://huggingface.co/AdithyaSK), presented at
|
| 16 |
+
**[AMD AI Dev Day](https://amd.indiadevday.com/)**.
|
| 17 |
+
|
| 18 |
+
RL environments (OpenEnv) and RL training (TRL) — what an environment actually is,
|
| 19 |
+
how reward hacking happens, and how to build and train against your own.
|
| 20 |
+
|
| 21 |
+
React + Vite. Arrow keys / clicker / swipe to navigate, `t` for light/dark, `f` for fullscreen.
|
| 22 |
+
|
| 23 |
+
Source: [adithya-s-k/RL_Envs_101](https://github.com/adithya-s-k/RL_Envs_101) →
|
| 24 |
+
`tutorials/slides/rl-environments-101-amd/`
|
assets/adithyask-CfgQu2fr.jpeg
ADDED
|
assets/blog-hero-A-OkyEfb.png
ADDED
|
Git LFS Details
|
assets/cheating-agents-Bpt4f8GA.jpeg
ADDED
|
assets/cognitivelab-BpHudwdd.png
ADDED
|
assets/index-BaIqJAjN.js
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
assets/index-DlQWIdL0.css
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
*,*:before,*:after{box-sizing:border-box}html,body,#root{margin:0;padding:0;width:100%;height:100%;overflow:hidden}body{font-family:ui-monospace,SF Mono,JetBrains Mono,Menlo,Monaco,Consolas,monospace;-webkit-font-smoothing:antialiased;text-rendering:optimizeLegibility}button{font-family:inherit;cursor:pointer}button:focus:not(:focus-visible){outline:none}
|
assets/openai-ss1-DOT2K7xh.png
ADDED
|
Git LFS Details
|
assets/openai-ss2-crop-BqCLkRBg.png
ADDED
|
assets/openai-ss3-crop-KZZxFTHs.png
ADDED
|
Git LFS Details
|
assets/qwen-scaling-C1rB1MJm.png
ADDED
|
Git LFS Details
|
assets/reward-hack-curves-Dsoa5-YV.png
ADDED
|
Git LFS Details
|
index.html
CHANGED
|
@@ -1,19 +1,43 @@
|
|
| 1 |
<!doctype html>
|
| 2 |
-
<html>
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
</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, maximum-scale=1.0, user-scalable=no" />
|
| 6 |
+
<title>Scaling RL for LLMs — Adithya S Kolavi</title>
|
| 7 |
+
<meta
|
| 8 |
+
name="description"
|
| 9 |
+
content="Scaling RL for LLMs: RL Environments and RL Training — a talk by Adithya S Kolavi at AMD AI Dev Day."
|
| 10 |
+
/>
|
| 11 |
+
<meta name="theme-color" content="#07090f" />
|
| 12 |
+
|
| 13 |
+
<!-- Social preview. og:image must be absolute, so it points at the deployed
|
| 14 |
+
Space; public/og.png is regenerated by scripts/make_og.py. -->
|
| 15 |
+
<meta property="og:type" content="website" />
|
| 16 |
+
<meta property="og:site_name" content="Scaling RL for LLMs" />
|
| 17 |
+
<meta property="og:url" content="https://adithyask-scaling-rl-for-llms-amd-ai-dev-day.static.hf.space/" />
|
| 18 |
+
<meta property="og:title" content="Scaling RL for LLMs — RL Environments and RL Training" />
|
| 19 |
+
<meta
|
| 20 |
+
property="og:description"
|
| 21 |
+
content="A talk by Adithya S Kolavi at AMD AI Dev Day: what an RL environment actually is, how reward hacking happens, and how to build and train your own with OpenEnv + TRL."
|
| 22 |
+
/>
|
| 23 |
+
<meta property="og:image" content="https://adithyask-scaling-rl-for-llms-amd-ai-dev-day.static.hf.space/og.png" />
|
| 24 |
+
<meta property="og:image:width" content="1200" />
|
| 25 |
+
<meta property="og:image:height" content="630" />
|
| 26 |
+
<meta property="og:image:alt" content="Scaling RL for LLMs — RL Environments and RL Training · Adithya S Kolavi" />
|
| 27 |
+
|
| 28 |
+
<meta name="twitter:card" content="summary_large_image" />
|
| 29 |
+
<meta name="twitter:site" content="@AdithyaSK" />
|
| 30 |
+
<meta name="twitter:creator" content="@AdithyaSK" />
|
| 31 |
+
<meta name="twitter:title" content="Scaling RL for LLMs — RL Environments and RL Training" />
|
| 32 |
+
<meta
|
| 33 |
+
name="twitter:description"
|
| 34 |
+
content="A talk by Adithya S Kolavi at AMD AI Dev Day: what an RL environment actually is, how reward hacking happens, and how to build and train your own with OpenEnv + TRL."
|
| 35 |
+
/>
|
| 36 |
+
<meta name="twitter:image" content="https://adithyask-scaling-rl-for-llms-amd-ai-dev-day.static.hf.space/og.png" />
|
| 37 |
+
<script type="module" crossorigin src="./assets/index-BaIqJAjN.js"></script>
|
| 38 |
+
<link rel="stylesheet" crossorigin href="./assets/index-DlQWIdL0.css">
|
| 39 |
+
</head>
|
| 40 |
+
<body>
|
| 41 |
+
<div id="root"></div>
|
| 42 |
+
</body>
|
| 43 |
</html>
|
og.png
ADDED
|