Spaces:
Running
Running
File size: 15,965 Bytes
381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a 381c120 aeed15a |
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 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 |
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description" content="Distributed Translation System for translating the DataTonic/dark_thoughts_case_study_merged dataset across multiple languages using RunPod and Ollama.">
<meta name="keywords" content="Distributed Translation, RunPod, Ollama, Dark Thoughts Dataset">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Distributed Translation System for Dark Thoughts Dataset</title>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro" rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="./static/css/fontawesome.all.min.css">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/images/favicon.svg">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script defer src="./static/js/fontawesome.all.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">Distributed Translation System for Dark Thoughts Dataset</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">Your Name or Team</span>
</div>
<div class="column has-text-centered">
<div class="publication-links">
<span class="link-block">
<a href="https://github.com/yourusername/distributed-translation" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon"><i class="fab fa-github"></i></span>
<span>Code</span>
</a>
</span>
<span class="link-block">
<a href="https://huggingface.co/datasets/DataTonic/dark_thoughts_case_study_merged" target="_blank" class="external-link button is-normal is-rounded is-dark">
<span class="icon"><i class="far fa-images"></i></span>
<span>Data</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3">Overview</h2>
<div class="content has-text-justified">
<p>
This project implements a distributed translation system using RunPod and Ollama to translate the <a href="https://huggingface.co/datasets/DataTonic/dark_thoughts_case_study_merged" target="_blank">DataTonic/dark_thoughts_case_study_merged</a> dataset across multiple languages. The system parses thinking content from responses and translates both components separately.
</p>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Architecture</h2>
<div class="content has-text-justified">
<p>The system consists of several components:</p>
<ol>
<li><strong>RunPod API Client</strong> (<code>runpodapi.py</code>): Handles communication with the RunPod API for creating, managing, and monitoring pods.</li>
<li><strong>RunPod Command Executor</strong> (<code>runcommandsrunpod.py</code>): Executes commands on RunPod instances and checks their readiness.</li>
<li><strong>RunPod Launcher</strong> (<code>runpodlauncher.py</code>): Manages the launching and coordination of multiple RunPod instances.</li>
<li><strong>RunPod Manager</strong> (<code>runpodmanager.py</code>): High-level manager for RunPod instances used for distributed translation.</li>
<li><strong>Ollama Client</strong> (<code>ollamaclient.py</code>): Async client for interacting with Ollama API and distributing translation tasks.</li>
<li><strong>Translation Coordinator</strong> (<code>translationcoordinator.py</code>): Orchestrates the translation process across dataset splits and languages.</li>
<li><strong>Data Processor</strong> (<code>dataprocessor.py</code>): Handles loading, processing, and saving the translated dataset.</li>
<li><strong>Main Script</strong> (<code>translate.py</code>): Entry point for running the distributed translation process.</li>
<li><strong>Test Scripts</strong> (<code>test_translation.py</code>, <code>test_parsing.py</code>): Tests the functionality of the distributed translation system.</li>
</ol>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Requirements</h2>
<div class="content has-text-justified">
<ul>
<li>Python 3.8+</li>
<li>RunPod API key</li>
<li>Access to RunPod GPU instances</li>
<li>The following Python packages: <code>aiohttp</code>, <code>asyncio</code>, <code>datasets</code>, <code>pandas</code>, <code>tqdm</code>, <code>requests</code>, <code>pydantic</code></li>
</ul>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Installation</h2>
<div class="content has-text-justified">
<ol>
<li>Clone the repository:
<pre><code>git clone https://github.com/yourusername/distributed-translation.git
cd distributed-translation</code></pre>
</li>
<li>Install the required packages:
<pre><code>pip install -r requirements.txt</code></pre>
</li>
<li>Set up your RunPod API key:
<pre><code>export RUNPOD_API_KEY=your_runpod_api_key</code></pre>
</li>
</ol>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Dataset Structure</h2>
<div class="content has-text-justified">
<p>The system works with the DataTonic/dark_thoughts_case_study_merged dataset, which contains:</p>
<ul>
<li>English split: 20,711 examples</li>
<li>Chinese split: 20,204 examples</li>
</ul>
<p>The system parses thinking content (text before <code></think></code>) from responses and translates both components separately.</p>
<p>The final dataset structure follows this model:</p>
<pre><code>class Feature(BaseModel):
id: int
thinking: str
response: str
thinking_translated: str
response_translated: str
query: str
source_data: str
category: str
endpoint: str
source: str</code></pre>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Usage</h2>
<h3 class="title is-4">Running the Translation Process</h3>
<div class="content has-text-justified">
<p>To run the full translation process:</p>
<pre><code>python translate.py --pod-count 40 --batch-size 16 --max-tokens 100</code></pre>
<p>Additional options:</p>
<pre><code>--api-key TEXT RunPod API key (defaults to RUNPOD_API_KEY environment variable)
--pod-count INTEGER Number of RunPod instances to launch (default: 40)
--dataset TEXT Dataset name or path (default: DataTonic/dark_thoughts_case_study_merged)
--output-dir TEXT Output directory for translated data (default: translated_dataset)
--batch-size INTEGER Batch size for translation (default: 16)
--max-tokens INTEGER Maximum number of tokens to generate (default: 100)
--gpu-type TEXT GPU type ID for RunPod instances (default: NVIDIA RTX A5000)
--image TEXT Docker image name (default: tonic01/ollama-gemmax2)
--model TEXT Model name for translation (default: gemmax2)
--cleanup Terminate all pods after completion
--prepare-only Only prepare the dataset without translating
--process-only Only process the translated dataset
--validate Validate dataset structure after processing</code></pre>
</div>
<h3 class="title is-4">Testing the System</h3>
<div class="content has-text-justified">
<p>To test the system components:</p>
<pre><code>python test_translation.py --test all</code></pre>
<p>To test the parsing functionality:</p>
<pre><code>python test_parsing.py --test all</code></pre>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Translation Process</h2>
<div class="content has-text-justified">
<p>The translation process follows these steps:</p>
<ol>
<li><strong>Preparation</strong>: Parse the dataset to separate thinking content from responses.</li>
<li><strong>Setup</strong>: Launch 40 RunPod instances with the <code>tonic01/ollama-gemmax2</code> Docker image.</li>
<li><strong>Readiness Check</strong>: Wait for all pods to be ready and for Ollama to be initialized with the required model.</li>
<li><strong>Translation</strong>:
<ul>
<li>For each dataset split (English and Chinese):</li>
<li>Translate thinking and response fields separately to all target languages.</li>
<li>Skip empty thinking content to optimize translation.</li>
<li>Save intermediate results periodically.</li>
</ul>
</li>
<li><strong>Processing</strong>: Merge translations and create a Hugging Face dataset structure.</li>
<li><strong>Validation</strong>: Ensure the dataset structure matches the required Feature model.</li>
<li><strong>Cleanup</strong>: Terminate all pods if requested.</li>
</ol>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Supported Languages</h2>
<div class="content has-text-justified">
<p>The system supports translation between the following languages:</p>
<p>Arabic, Bengali, Czech, German, English, Spanish, Persian, French, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Burmese, Dutch, Polish, Portuguese, Russian, Thai, Tagalog, Turkish, Urdu, Vietnamese, Chinese.</p>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Error Handling and Recovery</h2>
<div class="content has-text-justified">
<p>The system includes several error handling and recovery mechanisms:</p>
<ul>
<li><strong>Retry Logic</strong>: Failed translations are automatically retried.</li>
<li><strong>Checkpointing</strong>: Intermediate results are saved periodically to allow resuming from failures.</li>
<li><strong>Health Checks</strong>: Pod and Ollama health are checked before starting translation.</li>
<li><strong>Empty Content Handling</strong>: Empty thinking content is handled efficiently to avoid unnecessary translations.</li>
<li><strong>Graceful Termination</strong>: Resources are properly cleaned up on completion or failure.</li>
</ul>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Docker Image Requirements</h2>
<div class="content has-text-justified">
<p>The <code>tonic01/ollama-gemmax2</code> Docker image should have:</p>
<ol>
<li>Ollama installed and configured to run on port 11434</li>
<li>The GemmaX2-28-2B-v0.1 model pre-loaded or configured to load automatically</li>
<li>Sufficient GPU memory (at least 24GB recommended)</li>
</ol>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Example Workflow</h2>
<div class="content has-text-justified">
<ol>
<li><strong>Prepare Dataset</strong>:
<pre><code>python translate.py --prepare-only</code></pre>
</li>
<li><strong>Run Translation</strong>:
<pre><code>python translate.py --pod-count 40</code></pre>
</li>
<li><strong>Process Results Only</strong>:
<pre><code>python translate.py --process-only --validate</code></pre>
</li>
<li><strong>Cleanup</strong>:
<pre><code>python test_translation.py --test termination</code></pre>
</li>
</ol>
</div>
</div>
</div>
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3">Troubleshooting</h2>
<div class="content has-text-justified">
<ul>
<li><strong>API Key Issues</strong>: Ensure your RunPod API key is correctly set in the environment variable or passed as a parameter.</li>
<li><strong>GPU Availability</strong>: Check RunPod for GPU availability if pod creation fails.</li>
<li><strong>Model Loading</strong>: If Ollama readiness check times out, the model may be too large for the selected GPU type.</li>
<li><strong>Translation Errors</strong>: Check the logs for specific error messages. Most translation errors are automatically retried.</li>
<li><strong>Dataset Structure</strong>: Run with the <code>--validate</code> flag to ensure the dataset structure matches the required Feature model.</li>
</ul>
</div>
</div>
</div>
</div>
</section>
<section class="section" id="License">
<div class="container is-max-desktop content">
<h2 class="title">License</h2>
<div class="content has-text-justified">
<p>This project is licensed under the Apache 2.0 License - see the <a href="LICENSE" target="_blank">LICENSE</a> file for details.</p>
</div>
</div>
</section>
<footer class="footer">
<div class="container">
<div class="content has-text-centered">
<a class="icon-link" href="https://github.com/yourusername/distributed-translation" target="_blank">
<i class="fab fa-github"></i>
</a>
</div>
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p>
This website is licensed under a <a rel="license" href="http://creativecommons.org/licenses/by-sa/4.0/" target="_blank">Creative Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
<p>
This means you are free to borrow the <a href="https://github.com/yourusername/distributed-translation" target="_blank">source code</a> of this website, we just ask that you link back to this page in the footer.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html> |