{
  "id": 458552,
  "title": "201st Place Solution for the Google - Fast or Slow? Predict AI Model Runtime",
  "url": "/competitions/predict-ai-model-runtime/discussion/458552",
  "author_name": "",
  "post_date": "2023-11-30T14:09:16.320000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<h1>1. Introduction</h1>\n<p>I am happy to be part of Google - Fast or Slow? Predict AI Model Runtime. I want to express my appreciation to the organizers, sponsors, and Kaggle staff for their efforts, and I hope everyone has the best time. This competition and other participants provided me with a lot of knowledge.</p>\n<p>I am grateful to MIHU for providing the public notebook <a href=\"https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932\" target=\"_blank\">https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932</a></p>\n<h1>2. Context</h1>\n<ul>\n<li>Business context: <a href=\"https://www.kaggle.com/competitions/predict-ai-model-runtime\" target=\"_blank\">https://www.kaggle.com/competitions/predict-ai-model-runtime</a></li>\n<li>Data context: <a href=\"https://www.kaggle.com/competitions/predict-ai-model-runtime/data\" target=\"_blank\">https://www.kaggle.com/competitions/predict-ai-model-runtime/data</a></li>\n</ul>\n<h1>3. Overview of the approach</h1>\n<p>The solution was a copy of the public notebook(Public/Private LB of 0.39199/0.27376) <br>\nwith adding dataset created as output of public notebook (Public/Private LB of 0.39199/0.27376).</p>\n<p>The data preprocessing process: Adjacency matrix. Virtual first node.</p>\n<p>The algorithms employed: modified Bert.</p>\n<ul>\n<li>Inference:  <br>\n(copy codes given by MIHU in a public notebook  <a href=\"url\" target=\"_blank\">https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932</a>)<br>\nThe validation strategy: split train and valid dataset.</li>\n</ul>\n<h1>4. Method modified Bert.</h1>\n<p>Solution: </p>\n<table>\n<thead>\n<tr>\n<th>notebook</th>\n<th>score private</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>MIHU public <a href=\"https://www.kaggle.com/code/liudacheldieva/simple-prediction-for-five-datasets-2d0336?scriptVersionId=151176695\" target=\"_blank\">https://www.kaggle.com/code/liudacheldieva/simple-prediction-for-five-datasets-2d0336?scriptVersionId=151176695</a></td>\n<td>0.27376</td>\n</tr>\n<tr>\n<td>Change: add output of public notebook as input. Copy input to output.</td>\n<td>0.27376</td>\n</tr>\n</tbody>\n</table>\n<p><br></p>\n<h1>5. Final result</h1>\n<table>\n<thead>\n<tr>\n<th>public LB</th>\n<th>private LB</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td><strong>0.39199</strong></td>\n<td>0.27376</td>\n</tr>\n<tr>\n<td>0.13262</td>\n<td>0.15293</td>\n</tr>\n<tr>\n<td>0.12882</td>\n<td>0.14862</td>\n</tr>\n<tr>\n<td>0.14798</td>\n<td>0.12904</td>\n</tr>\n<tr>\n<td>0.14798</td>\n<td>0.12904</td>\n</tr>\n<tr>\n<td>0.15345</td>\n<td>0.13017</td>\n</tr>\n</tbody>\n</table>\n<h1>6. Sources</h1>\n<ul>\n<li><a href=\"url\" target=\"_blank\">https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932</a></li>\n<li><a href=\"url\" target=\"_blank\">https://www.kaggle.com/datasets/liudacheldieva/gf-sub-last</a></li>\n</ul>",
  "messages": [
    {
      "id": 2543990,
      "postDate": "2023-11-30T14:09:16.320Z",
      "content": "<h1>1. Introduction</h1>\n<p>I am happy to be part of Google - Fast or Slow? Predict AI Model Runtime. I want to express my appreciation to the organizers, sponsors, and Kaggle staff for their efforts, and I hope everyone has the best time. This competition and other participants provided me with a lot of knowledge.</p>\n<p>I am grateful to MIHU for providing the public notebook <a href=\"https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932\" target=\"_blank\">https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932</a></p>\n<h1>2. Context</h1>\n<ul>\n<li>Business context: <a href=\"https://www.kaggle.com/competitions/predict-ai-model-runtime\" target=\"_blank\">https://www.kaggle.com/competitions/predict-ai-model-runtime</a></li>\n<li>Data context: <a href=\"https://www.kaggle.com/competitions/predict-ai-model-runtime/data\" target=\"_blank\">https://www.kaggle.com/competitions/predict-ai-model-runtime/data</a></li>\n</ul>\n<h1>3. Overview of the approach</h1>\n<p>The solution was a copy of the public notebook(Public/Private LB of 0.39199/0.27376) <br>\nwith adding dataset created as output of public notebook (Public/Private LB of 0.39199/0.27376).</p>\n<p>The data preprocessing process: Adjacency matrix. Virtual first node.</p>\n<p>The algorithms employed: modified Bert.</p>\n<ul>\n<li>Inference:  <br>\n(copy codes given by MIHU in a public notebook  <a href=\"url\" target=\"_blank\">https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932</a>)<br>\nThe validation strategy: split train and valid dataset.</li>\n</ul>\n<h1>4. Method modified Bert.</h1>\n<p>Solution: </p>\n<table>\n<thead>\n<tr>\n<th>notebook</th>\n<th>score private</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>MIHU public <a href=\"https://www.kaggle.com/code/liudacheldieva/simple-prediction-for-five-datasets-2d0336?scriptVersionId=151176695\" target=\"_blank\">https://www.kaggle.com/code/liudacheldieva/simple-prediction-for-five-datasets-2d0336?scriptVersionId=151176695</a></td>\n<td>0.27376</td>\n</tr>\n<tr>\n<td>Change: add output of public notebook as input. Copy input to output.</td>\n<td>0.27376</td>\n</tr>\n</tbody>\n</table>\n<p><br></p>\n<h1>5. Final result</h1>\n<table>\n<thead>\n<tr>\n<th>public LB</th>\n<th>private LB</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td><strong>0.39199</strong></td>\n<td>0.27376</td>\n</tr>\n<tr>\n<td>0.13262</td>\n<td>0.15293</td>\n</tr>\n<tr>\n<td>0.12882</td>\n<td>0.14862</td>\n</tr>\n<tr>\n<td>0.14798</td>\n<td>0.12904</td>\n</tr>\n<tr>\n<td>0.14798</td>\n<td>0.12904</td>\n</tr>\n<tr>\n<td>0.15345</td>\n<td>0.13017</td>\n</tr>\n</tbody>\n</table>\n<h1>6. Sources</h1>\n<ul>\n<li><a href=\"url\" target=\"_blank\">https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932</a></li>\n<li><a href=\"url\" target=\"_blank\">https://www.kaggle.com/datasets/liudacheldieva/gf-sub-last</a></li>\n</ul>",
      "rawMarkdown": "# 1. Introduction\n\nI am happy to be part of Google - Fast or Slow? Predict AI Model Runtime. I want to express my appreciation to the organizers, sponsors, and Kaggle staff for their efforts, and I hope everyone has the best time. This competition and other participants provided me with a lot of knowledge.\n\nI am grateful to MIHU for providing the public notebook https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932\n\n# 2. Context\n\n- Business context: https://www.kaggle.com/competitions/predict-ai-model-runtime\n- Data context: https://www.kaggle.com/competitions/predict-ai-model-runtime/data\n\n# 3. Overview of the approach\n\nThe solution was a copy of the public notebook(Public/Private LB of 0.39199/0.27376) \nwith adding dataset created as output of public notebook (Public/Private LB of 0.39199/0.27376).\n \nThe data preprocessing process: Adjacency matrix. Virtual first node.\n\nThe algorithms employed: modified Bert.\n- Inference:  \n(copy codes given by MIHU in a public notebook  [https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932](url))\nThe validation strategy: split train and valid dataset.\n\n# 4. Method modified Bert.\n\nSolution: \n\n| notebook | score private | \n| --- | --- |\n| MIHU public https://www.kaggle.com/code/liudacheldieva/simple-prediction-for-five-datasets-2d0336?scriptVersionId=151176695 | 0.27376 | \n| Change: add output of public notebook as input. Copy input to output. |  0.27376  |\n\n<br>\n# 5. Final result\n\n |  public LB | private LB |\n | --- | --- | \n | **0.39199**    | 0.27376 |\n |  0.13262 | 0.15293 |\n | 0.12882   | 0.14862|\n |  0.14798 | 0.12904 | \n |  0.14798 | 0.12904 | \n | 0.15345 | 0.13017 | \n\n\n# 6. Sources\n- [https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932](url)\n- [https://www.kaggle.com/datasets/liudacheldieva/gf-sub-last](url)",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2543990": "# 1. Introduction\n\nI am happy to be part of Google - Fast or Slow? Predict AI Model Runtime. I want to express my appreciation to the organizers, sponsors, and Kaggle staff for their efforts, and I hope everyone has the best time. This competition and other participants provided me with a lot of knowledge.\n\nI am grateful to MIHU for providing the public notebook https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932\n\n# 2. Context\n\n- Business context: https://www.kaggle.com/competitions/predict-ai-model-runtime\n- Data context: https://www.kaggle.com/competitions/predict-ai-model-runtime/data\n\n# 3. Overview of the approach\n\nThe solution was a copy of the public notebook(Public/Private LB of 0.39199/0.27376) \nwith adding dataset created as output of public notebook (Public/Private LB of 0.39199/0.27376).\n \nThe data preprocessing process: Adjacency matrix. Virtual first node.\n\nThe algorithms employed: modified Bert.\n- Inference:  \n(copy codes given by MIHU in a public notebook  [https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932](url))\nThe validation strategy: split train and valid dataset.\n\n# 4. Method modified Bert.\n\nSolution: \n\n| notebook | score private | \n| --- | --- |\n| MIHU public https://www.kaggle.com/code/liudacheldieva/simple-prediction-for-five-datasets-2d0336?scriptVersionId=151176695 | 0.27376 | \n| Change: add output of public notebook as input. Copy input to output. |  0.27376  |\n\n<br>\n# 5. Final result\n\n |  public LB | private LB |\n | --- | --- | \n | **0.39199**    | 0.27376 |\n |  0.13262 | 0.15293 |\n | 0.12882   | 0.14862|\n |  0.14798 | 0.12904 | \n |  0.14798 | 0.12904 | \n | 0.15345 | 0.13017 | \n\n\n# 6. Sources\n- [https://www.kaggle.com/code/chenboluo/simple-prediction-for-five-datasets?scriptVersionId=145767932](url)\n- [https://www.kaggle.com/datasets/liudacheldieva/gf-sub-last](url)"
  }
}