{
  "id": 242704,
  "title": "Can we train a model outside with given data and attach it as a pretrain model and do inference on test data?",
  "url": "/competitions/coleridgeinitiative-show-us-the-data/discussion/242704",
  "author_name": "",
  "post_date": "2021-05-30T10:05:20.155755500Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi Guys,</p>\n<p>Been some time since I have been to Kaggle.<br>\nSo just wanted to clarify, can we download the train data and train it outside and then use that model as pre-trained model to make inference on test data?</p>\n<p>Sorry, if its a blunder question. Just to clarify .</p>",
  "messages": [
    {
      "id": "1328507",
      "postDate": "05/30/2021 10:05:20",
      "content": "<p>Hi Guys,</p>\n<p>Been some time since I have been to Kaggle.<br>\nSo just wanted to clarify, can we download the train data and train it outside and then use that model as pre-trained model to make inference on test data?</p>\n<p>Sorry, if its a blunder question. Just to clarify .</p>",
      "rawMarkdown": "Hi Guys,\n\nBeen some time since I have been to Kaggle.\nSo just wanted to clarify, can we download the train data and train it outside and then use that model as pre-trained model to make inference on test data?\n\nSorry, if its a blunder question. Just to clarify .",
      "votes": null
    },
    {
      "id": "1328537",
      "postDate": "05/30/2021 10:37:59",
      "content": "<p>Yes, you can create an inference-only kernel with a pre-train model.</p>\n<p>You don't have to train a model on a kaggle kernel.</p>",
      "rawMarkdown": "Yes, you can create an inference-only kernel with a pre-train model.\n\nYou don't have to train a model on a kaggle kernel.",
      "votes": null
    },
    {
      "id": "1328542",
      "postDate": "05/30/2021 10:42:33",
      "content": "<p>Thanks a lot</p>",
      "rawMarkdown": "Thanks a lot",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1328537,
      "author_name": "spshota",
      "author_url": "",
      "post_date": "05/30/2021 10:37:59",
      "content": "<p>Yes, you can create an inference-only kernel with a pre-train model.</p>\n<p>You don't have to train a model on a kaggle kernel.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1328542,
      "author_name": "s4sarath",
      "author_url": "",
      "post_date": "05/30/2021 10:42:33",
      "content": "<p>Thanks a lot</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1328507": "Hi Guys,\n\nBeen some time since I have been to Kaggle.\nSo just wanted to clarify, can we download the train data and train it outside and then use that model as pre-trained model to make inference on test data?\n\nSorry, if its a blunder question. Just to clarify .",
    "1328537": "Yes, you can create an inference-only kernel with a pre-train model.\n\nYou don't have to train a model on a kaggle kernel.",
    "1328542": "Thanks a lot"
  },
  "source": "meta"
}