{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreicwrahzyvar5r2r7livazcxx2fcyoe5ggubbz52v5qgjzeckh2c5m",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3ml4t42admw22"
  },
  "path": "/t/feature-request-kernels-metadata-in-the-models/175769#post_1",
  "publishedAt": "2026-05-05T18:28:12.000Z",
  "site": "https://discuss.huggingface.co",
  "textContent": "# Feature Request: Kernel Metadata for Models\n\nWhen Kernels were introduced, this idea immediately came to mind: being able to link the kernels used in a model repo, in a way very similar to how datasets are linked.\nI imagine they might look like this:\n\n\nAnd the YAML code could then be like this, for example:\n\n\n    ---\n    kernels:\n      - kernels-community/flash-attn3\n    ---\n    # Model Card\n\n\n### Advantages\n\nThe advantages are:\n\n  * a user can immediately know whether their GPU is compatible or not for training\n  * Many users use tags to express the kernel, but this can create confusion. Some might use “flash-attention” or “flash_attn.” Having the ability to directly insert the kernel solves this problem.\n\n\n\nI hope you like it and that the Hugging Face team might consider implementing it.\nThanks for have reading",
  "title": "Feature request: Kernels metadata in the models"
}