{
  "id": 667105,
  "title": "Conditioning on species",
  "url": "/competitions/csiro-biomass/discussion/667105",
  "author_name": "Ali Ozan Memetoglu",
  "post_date": "2026-01-11T04:07:17.623000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello since we have the same species in the test set, has anyone tried an architecture like classifying the image first and conditioning the model on it? But also since we have very limited training data, an auxiliary loss might make more sense.</p>",
  "messages": [
    {
      "id": 3389398,
      "postDate": "2026-01-11T04:07:17.623Z",
      "content": "<p>Hello since we have the same species in the test set, has anyone tried an architecture like classifying the image first and conditioning the model on it? But also since we have very limited training data, an auxiliary loss might make more sense.</p>",
      "rawMarkdown": "Hello since we have the same species in the test set, has anyone tried an architecture like classifying the image first and conditioning the model on it? But also since we have very limited training data, an auxiliary loss might make more sense.",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "3389398": "Hello since we have the same species in the test set, has anyone tried an architecture like classifying the image first and conditioning the model on it? But also since we have very limited training data, an auxiliary loss might make more sense."
  }
}