{
  "id": 414311,
  "title": "Clarification in License",
  "url": "/competitions/benetech-making-graphs-accessible/discussion/414311",
  "author_name": "",
  "post_date": "2023-06-01T07:32:47.176015500Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I have come across datasets and models with licenses other than MIT, but they do support commercial usage. Am I allowed to use both the datasets and models? Could you clarify which of the following licenses can be used: 1. Apache-2.0 license 2. GPL-3.0 license 3. CC0: Public Domain? <br>\nAnd is the MIT License condition valid only for winners or for all Kagglers submitting their work?<br>\n<a href=\"https://www.kaggle.com/ryanholbrook\" target=\"_blank\">@ryanholbrook</a> <a href=\"https://www.kaggle.com/maggiemd\" target=\"_blank\">@maggiemd</a></p>",
  "messages": [
    {
      "id": "2283297",
      "postDate": "06/01/2023 07:32:47",
      "content": "<p>I have come across datasets and models with licenses other than MIT, but they do support commercial usage. Am I allowed to use both the datasets and models? Could you clarify which of the following licenses can be used: 1. Apache-2.0 license 2. GPL-3.0 license 3. CC0: Public Domain? <br>\nAnd is the MIT License condition valid only for winners or for all Kagglers submitting their work?<br>\n<a href=\"https://www.kaggle.com/ryanholbrook\" target=\"_blank\">@ryanholbrook</a> <a href=\"https://www.kaggle.com/maggiemd\" target=\"_blank\">@maggiemd</a></p>",
      "rawMarkdown": "I have come across datasets and models with licenses other than MIT, but they do support commercial usage. Am I allowed to use both the datasets and models? Could you clarify which of the following licenses can be used: 1. Apache-2.0 license 2. GPL-3.0 license 3. CC0: Public Domain? \nAnd is the MIT License condition valid only for winners or for all Kagglers submitting their work?\n@ryanholbrook @maggiemd",
      "votes": null
    },
    {
      "id": "2283470",
      "postDate": "06/01/2023 09:14:49",
      "content": "<p>This is just my personal opinion,</p>\n<ul>\n<li><p>Apache-2.0, CC0<br>\nThese should be allowed.<br>\nThese are non-copylefted, and Apache-2.0 is almost as restrictive as MIT, CC0, since they are almost in the public domain with the copyright abandoned, I don't think there is any problem in using them.<br>\nIncidentally, TensorFlow, OpenCV, and transformers etc. should be released under the Apace 2.0 license, if I recall correctly, If these licenses were restricted, it would mean that almost all solutions would not be allowed.</p></li>\n<li><p>GPL-3.0<br>\nCertainly commercial use is possible, but probably not allowed due to strong restrictions such as copyleft and source code disclosure obligations.<br>\n(Though ultimately it is up to the organizer to decide.)<br>\nIn fact, in some past competitions, the GPL license was regarded as effectively restricting commercial use, and use was not permitted.</p></li>\n<li><p>Does the MIT License condition apply only to the winner?<br>\nI think you would have to be the organizer to know this.<br>\nBut realistically it is impossible to check all the submissions, so I don't think you can find out if anyone other than the winners probably used a license that is not allowed.<br>\nBut in my personal opinion, I think all participants have equal chances to win prizes, and basically I think all participants should follow the licensing rules.</p></li>\n</ul>\n<p>However, again, please keep in mind that these are just my personal opinions.<br>\nI am not a legal expert, nor am I the organizer.<br>\nI think the only way to get a formal answer is to have the organizer give it to me.</p>",
      "rawMarkdown": "This is just my personal opinion,\n\n-  Apache-2.0, CC0\nThese should be allowed.\nThese are non-copylefted, and Apache-2.0 is almost as restrictive as MIT, CC0, since they are almost in the public domain with the copyright abandoned, I don't think there is any problem in using them.\nIncidentally, TensorFlow, OpenCV, and transformers etc. should be released under the Apace 2.0 license, if I recall correctly, If these licenses were restricted, it would mean that almost all solutions would not be allowed.\n\n- GPL-3.0\nCertainly commercial use is possible, but probably not allowed due to strong restrictions such as copyleft and source code disclosure obligations.\n(Though ultimately it is up to the organizer to decide.)\nIn fact, in some past competitions, the GPL license was regarded as effectively restricting commercial use, and use was not permitted.\n\n- Does the MIT License condition apply only to the winner?\nI think you would have to be the organizer to know this.\nBut realistically it is impossible to check all the submissions, so I don't think you can find out if anyone other than the winners probably used a license that is not allowed.\nBut in my personal opinion, I think all participants have equal chances to win prizes, and basically I think all participants should follow the licensing rules.\n\n\nHowever, again, please keep in mind that these are just my personal opinions.\nI am not a legal expert, nor am I the organizer.\nI think the only way to get a formal answer is to have the organizer give it to me.",
      "votes": null
    },
    {
      "id": "2283552",
      "postDate": "06/01/2023 10:43:21",
      "content": "<p>Thank a lot. Yeah you opinion sounds correct. I have the same view on it.<br>\nOne more question.<br>\nFor this problem, Deplot / Matcha/ Donut which gives good accuracy with robustness?<br>\nCould you please tell which is better!</p>",
      "rawMarkdown": "Thank a lot. Yeah you opinion sounds correct. I have the same view on it.\nOne more question.\nFor this problem, Deplot / Matcha/ Donut which gives good accuracy with robustness?\nCould you please tell which is better!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2283470,
      "author_name": "kashiwaba",
      "author_url": "",
      "post_date": "06/01/2023 09:14:49",
      "content": "<p>This is just my personal opinion,</p>\n<ul>\n<li><p>Apache-2.0, CC0<br>\nThese should be allowed.<br>\nThese are non-copylefted, and Apache-2.0 is almost as restrictive as MIT, CC0, since they are almost in the public domain with the copyright abandoned, I don't think there is any problem in using them.<br>\nIncidentally, TensorFlow, OpenCV, and transformers etc. should be released under the Apace 2.0 license, if I recall correctly, If these licenses were restricted, it would mean that almost all solutions would not be allowed.</p></li>\n<li><p>GPL-3.0<br>\nCertainly commercial use is possible, but probably not allowed due to strong restrictions such as copyleft and source code disclosure obligations.<br>\n(Though ultimately it is up to the organizer to decide.)<br>\nIn fact, in some past competitions, the GPL license was regarded as effectively restricting commercial use, and use was not permitted.</p></li>\n<li><p>Does the MIT License condition apply only to the winner?<br>\nI think you would have to be the organizer to know this.<br>\nBut realistically it is impossible to check all the submissions, so I don't think you can find out if anyone other than the winners probably used a license that is not allowed.<br>\nBut in my personal opinion, I think all participants have equal chances to win prizes, and basically I think all participants should follow the licensing rules.</p></li>\n</ul>\n<p>However, again, please keep in mind that these are just my personal opinions.<br>\nI am not a legal expert, nor am I the organizer.<br>\nI think the only way to get a formal answer is to have the organizer give it to me.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2283552,
          "author_name": "navinkumarmnk",
          "author_url": "",
          "post_date": "06/01/2023 10:43:21",
          "content": "<p>Thank a lot. Yeah you opinion sounds correct. I have the same view on it.<br>\nOne more question.<br>\nFor this problem, Deplot / Matcha/ Donut which gives good accuracy with robustness?<br>\nCould you please tell which is better!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2283297": "I have come across datasets and models with licenses other than MIT, but they do support commercial usage. Am I allowed to use both the datasets and models? Could you clarify which of the following licenses can be used: 1. Apache-2.0 license 2. GPL-3.0 license 3. CC0: Public Domain? \nAnd is the MIT License condition valid only for winners or for all Kagglers submitting their work?\n@ryanholbrook @maggiemd",
    "2283470": "This is just my personal opinion,\n\n-  Apache-2.0, CC0\nThese should be allowed.\nThese are non-copylefted, and Apache-2.0 is almost as restrictive as MIT, CC0, since they are almost in the public domain with the copyright abandoned, I don't think there is any problem in using them.\nIncidentally, TensorFlow, OpenCV, and transformers etc. should be released under the Apace 2.0 license, if I recall correctly, If these licenses were restricted, it would mean that almost all solutions would not be allowed.\n\n- GPL-3.0\nCertainly commercial use is possible, but probably not allowed due to strong restrictions such as copyleft and source code disclosure obligations.\n(Though ultimately it is up to the organizer to decide.)\nIn fact, in some past competitions, the GPL license was regarded as effectively restricting commercial use, and use was not permitted.\n\n- Does the MIT License condition apply only to the winner?\nI think you would have to be the organizer to know this.\nBut realistically it is impossible to check all the submissions, so I don't think you can find out if anyone other than the winners probably used a license that is not allowed.\nBut in my personal opinion, I think all participants have equal chances to win prizes, and basically I think all participants should follow the licensing rules.\n\n\nHowever, again, please keep in mind that these are just my personal opinions.\nI am not a legal expert, nor am I the organizer.\nI think the only way to get a formal answer is to have the organizer give it to me.",
    "2283552": "Thank a lot. Yeah you opinion sounds correct. I have the same view on it.\nOne more question.\nFor this problem, Deplot / Matcha/ Donut which gives good accuracy with robustness?\nCould you please tell which is better!"
  },
  "source": "meta"
}