{
  "id": 456450,
  "title": "17th place solution: GNN + TransformerEncoderLayer w/ BCELoss.",
  "url": "/competitions/predict-ai-model-runtime/discussion/456450",
  "author_name": "T0m",
  "post_date": "2023-11-20T05:55:37.747000",
  "votes": 8,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Thanks for hosting the interesting competition and congratulations to the winners!<br>\nI was disappointed with the result, but I learned a lot.</p>\n<p>only describe T0m part.</p>\n<h1>Layout</h1>\n<ul>\n<li>GNN(SAGEConv) + TransformerEncoderLayer</li>\n<li>BCELoss</li>\n<li>Predictions were made for all pairs, and the average of the predictions was used as a final guideline for sorting</li>\n<li>For tests with graphs similar to train, Lightgbm was used to train and make predictions individually. Only node_config_feat was used in that case.<ul>\n<li>This is not the cause of the shake down.</li></ul></li>\n</ul>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4705999%2F6520e81b8d8ee65c6b97dbf86ec6f5ce%2F2023-11-20%2014.48.00.png?generation=1700459296767515&amp;alt=media\" alt=\"\"></p>",
  "messages": [
    {
      "id": 2531328,
      "postDate": "2023-11-20T05:55:37.747Z",
      "content": "<p>Thanks for hosting the interesting competition and congratulations to the winners!<br>\nI was disappointed with the result, but I learned a lot.</p>\n<p>only describe T0m part.</p>\n<h1>Layout</h1>\n<ul>\n<li>GNN(SAGEConv) + TransformerEncoderLayer</li>\n<li>BCELoss</li>\n<li>Predictions were made for all pairs, and the average of the predictions was used as a final guideline for sorting</li>\n<li>For tests with graphs similar to train, Lightgbm was used to train and make predictions individually. Only node_config_feat was used in that case.<ul>\n<li>This is not the cause of the shake down.</li></ul></li>\n</ul>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4705999%2F6520e81b8d8ee65c6b97dbf86ec6f5ce%2F2023-11-20%2014.48.00.png?generation=1700459296767515&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "Thanks for hosting the interesting competition and congratulations to the winners!\nI was disappointed with the result, but I learned a lot.\n\nonly describe T0m part.\n\n# Layout\n\n- GNN(SAGEConv) + TransformerEncoderLayer\n- BCELoss\n- Predictions were made for all pairs, and the average of the predictions was used as a final guideline for sorting\n- For tests with graphs similar to train, Lightgbm was used to train and make predictions individually. Only node_config_feat was used in that case.\n  - This is not the cause of the shake down.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4705999%2F6520e81b8d8ee65c6b97dbf86ec6f5ce%2F2023-11-20%2014.48.00.png?generation=1700459296767515&alt=media)\n",
      "votes": 8
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2531328": "Thanks for hosting the interesting competition and congratulations to the winners!\nI was disappointed with the result, but I learned a lot.\n\nonly describe T0m part.\n\n# Layout\n\n- GNN(SAGEConv) + TransformerEncoderLayer\n- BCELoss\n- Predictions were made for all pairs, and the average of the predictions was used as a final guideline for sorting\n- For tests with graphs similar to train, Lightgbm was used to train and make predictions individually. Only node_config_feat was used in that case.\n  - This is not the cause of the shake down.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4705999%2F6520e81b8d8ee65c6b97dbf86ec6f5ce%2F2023-11-20%2014.48.00.png?generation=1700459296767515&alt=media)\n"
  }
}