{
  "id": 171176,
  "title": "Internet connection and Pretrained weights",
  "url": "/competitions/landmark-retrieval-2020/discussion/171176",
  "author_name": "",
  "post_date": "2020-07-30T17:41:23.114000600Z",
  "votes": 5,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hello, I'm new to Kaggle (and also machine learning) so I wanted to ask something about the rules.\nIt's not possible to submit a notebook that's using internet and in this way I'm not able to use pre-trained weights (ex: ResNet-101 trained on imagenet)  but to train a network from scratch takes more than 9 hours I guess... so am I allowed to upload weights manually to the Kaggle notebook?</p>",
  "messages": [
    {
      "id": "952189",
      "postDate": "07/30/2020 17:41:23",
      "content": "<p>Hello, I'm new to Kaggle (and also machine learning) so I wanted to ask something about the rules.\nIt's not possible to submit a notebook that's using internet and in this way I'm not able to use pre-trained weights (ex: ResNet-101 trained on imagenet)  but to train a network from scratch takes more than 9 hours I guess... so am I allowed to upload weights manually to the Kaggle notebook?</p>",
      "rawMarkdown": "Hello, I'm new to Kaggle (and also machine learning) so I wanted to ask something about the rules.\nIt's not possible to submit a notebook that's using internet and in this way I'm not able to use pre-trained weights (ex: ResNet-101 trained on imagenet)  but to train a network from scratch takes more than 9 hours I guess... so am I allowed to upload weights manually to the Kaggle notebook?",
      "votes": null
    },
    {
      "id": "952367",
      "postDate": "07/30/2020 21:09:17",
      "content": "<p>Hello, yes, you can upload them the same as a dataset from your notebook</p>",
      "rawMarkdown": "Hello, yes, you can upload them the same as a dataset from your notebook",
      "votes": null
    },
    {
      "id": "952420",
      "postDate": "07/30/2020 22:33:42",
      "content": "<p>Yep! You can add a dataset.</p>",
      "rawMarkdown": "Yep! You can add a dataset.",
      "votes": null
    },
    {
      "id": "952523",
      "postDate": "07/31/2020 02:34:48",
      "content": "<p>Yes, the use of weights that were trained at local machine is allowed.\nso are you going to download the dataset nd then upload weighs of trained model?\nalso\nYou may use data other than the Competition Data (“External Data”) to develop and test your models and Submissions. However, you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.</p>",
      "rawMarkdown": "Yes, the use of weights that were trained at local machine is allowed.\nso are you going to download the dataset nd then upload weighs of trained model?\nalso\nYou may use data other than the Competition Data (“External Data”) to develop and test your models and Submissions. However, you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 952367,
      "author_name": "fzovpec2",
      "author_url": "",
      "post_date": "07/30/2020 21:09:17",
      "content": "<p>Hello, yes, you can upload them the same as a dataset from your notebook</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 952420,
      "author_name": "camaskew",
      "author_url": "",
      "post_date": "07/30/2020 22:33:42",
      "content": "<p>Yep! You can add a dataset.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 952523,
      "author_name": "mohitduklan",
      "author_url": "",
      "post_date": "07/31/2020 02:34:48",
      "content": "<p>Yes, the use of weights that were trained at local machine is allowed.\nso are you going to download the dataset nd then upload weighs of trained model?\nalso\nYou may use data other than the Competition Data (“External Data”) to develop and test your models and Submissions. However, you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "952189": "Hello, I'm new to Kaggle (and also machine learning) so I wanted to ask something about the rules.\nIt's not possible to submit a notebook that's using internet and in this way I'm not able to use pre-trained weights (ex: ResNet-101 trained on imagenet)  but to train a network from scratch takes more than 9 hours I guess... so am I allowed to upload weights manually to the Kaggle notebook?",
    "952367": "Hello, yes, you can upload them the same as a dataset from your notebook",
    "952420": "Yep! You can add a dataset.",
    "952523": "Yes, the use of weights that were trained at local machine is allowed.\nso are you going to download the dataset nd then upload weighs of trained model?\nalso\nYou may use data other than the Competition Data (“External Data”) to develop and test your models and Submissions. However, you will (i) ensure the External Data is available to use by all participants of the competition for purposes of the competition at no cost to the other participants and (ii) post such access to the External Data for the participants to the official competition forum prior to the Entry Deadline."
  },
  "source": "meta"
}