{
  "id": 458329,
  "title": "84th Solution: Based on Transformer, the GraphTPU model",
  "url": "/competitions/predict-ai-model-runtime/discussion/458329",
  "author_name": "Haoqin Hong",
  "post_date": "2023-11-29T09:57:36.338000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>In this task, we enhanced the accuracy of the model by replacing the MLP layout in the open-source code Google GraphTPU with a Transformer. Additionally, we employed preprocessing techniques that effectively contributed to the successful completion of the task.</p>",
  "messages": [
    {
      "id": 2542495,
      "postDate": "2023-11-29T09:57:36.337Z",
      "content": "<p>In this task, we enhanced the accuracy of the model by replacing the MLP layout in the open-source code Google GraphTPU with a Transformer. Additionally, we employed preprocessing techniques that effectively contributed to the successful completion of the task.</p>",
      "rawMarkdown": "In this task, we enhanced the accuracy of the model by replacing the MLP layout in the open-source code Google GraphTPU with a Transformer. Additionally, we employed preprocessing techniques that effectively contributed to the successful completion of the task."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2542495": "In this task, we enhanced the accuracy of the model by replacing the MLP layout in the open-source code Google GraphTPU with a Transformer. Additionally, we employed preprocessing techniques that effectively contributed to the successful completion of the task."
  }
}