{
  "id": 77460,
  "title": "Does kaggle offer (free) computing?",
  "url": "/competitions/LANL-Earthquake-Prediction/discussion/77460",
  "author_name": "",
  "post_date": "2019-01-13T02:21:13.321032900Z",
  "votes": null,
  "comment_count": 8,
  "views": 0,
  "content": "<p>Just tried to build a earthquake prediction model and soon realized that my personal computer does not have enough resources to do so. If kaggle offers computer nodes for us to upload code and run the hosted data, it would be great.</p>\n\n<p>Thanks.</p>",
  "messages": [
    {
      "id": "455114",
      "postDate": "01/13/2019 02:21:13",
      "content": "<p>Just tried to build a earthquake prediction model and soon realized that my personal computer does not have enough resources to do so. If kaggle offers computer nodes for us to upload code and run the hosted data, it would be great.</p>\n\n<p>Thanks.</p>",
      "rawMarkdown": "Just tried to build a earthquake prediction model and soon realized that my personal computer does not have enough resources to do so. If kaggle offers computer nodes for us to upload code and run the hosted data, it would be great.\n\nThanks.",
      "votes": null
    },
    {
      "id": "455119",
      "postDate": "01/13/2019 02:49:49",
      "content": "<p>Yes, just go in kernels and 'New Kernel'. You can use a notebook or script environment (in the browser) and the machine has 16gb of ram, which is probably enough for this competition.</p>",
      "rawMarkdown": "Yes, just go in kernels and 'New Kernel'. You can use a notebook or script environment (in the browser) and the machine has 16gb of ram, which is probably enough for this competition.",
      "votes": null
    },
    {
      "id": "455145",
      "postDate": "01/13/2019 04:19:18",
      "content": "<p>You can also use Google Colaboratory, with access to TPU\n<a href=\"https://colab.research.google.com/notebooks/welcome.ipynb\">https://colab.research.google.com/notebooks/welcome.ipynb</a></p>",
      "rawMarkdown": "You can also use Google Colaboratory, with access to TPU\nhttps://colab.research.google.com/notebooks/welcome.ipynb",
      "votes": null
    },
    {
      "id": "455539",
      "postDate": "01/14/2019 06:14:33",
      "content": "<p>Yes they offer. That is enough for competition.</p>",
      "rawMarkdown": "Yes they offer. That is enough for competition.",
      "votes": null
    },
    {
      "id": "456066",
      "postDate": "01/15/2019 03:59:11",
      "content": "<p>Yes kaggle offers enough computing resources which include 16gb ram and you can also access gpu too when you create a new kernel by clicking new kernel button in the kernels page . Also you can upload your own data in to the kaggle kernel </p>",
      "rawMarkdown": "Yes kaggle offers enough computing resources which include 16gb ram and you can also access gpu too when you create a new kernel by clicking new kernel button in the kernels page . Also you can upload your own data in to the kaggle kernel",
      "votes": null
    },
    {
      "id": "456548",
      "postDate": "01/16/2019 03:06:06",
      "content": "<p>(deleted. see below)</p>",
      "rawMarkdown": "(deleted. see below)",
      "votes": null
    },
    {
      "id": "456550",
      "postDate": "01/16/2019 03:07:21",
      "content": "<p>Guys, thanks a lot for the info. Good to know kaggle offers the resources. The training data alone is 9.5GB. Loading them into memory, plus the N (look back time steps) * 9.5GB to feed into LSTM, I am afraid 16GB might not be enough? One has to do something smart to reduce the memory usage, I guess.</p>",
      "rawMarkdown": "Guys, thanks a lot for the info. Good to know kaggle offers the resources. The training data alone is 9.5GB. Loading them into memory, plus the N (look back time steps) * 9.5GB to feed into LSTM, I am afraid 16GB might not be enough? One has to do something smart to reduce the memory usage, I guess.",
      "votes": null
    },
    {
      "id": "456585",
      "postDate": "01/16/2019 05:43:19",
      "content": "<p>Hey please check this kernel to know about reducing the memory usage when loading the data into dataframe <a href=\"https://www.kaggle.com/gemartin/load-data-reduce-memory-usage\">https://www.kaggle.com/gemartin/load-data-reduce-memory-usage</a></p>",
      "rawMarkdown": "Hey please check this kernel to know about reducing the memory usage when loading the data into dataframe https://www.kaggle.com/gemartin/load-data-reduce-memory-usage",
      "votes": null
    },
    {
      "id": "456681",
      "postDate": "01/16/2019 09:38:44",
      "content": "<p>i am working on emoji expression recognization.\ncan anyone help me?</p>",
      "rawMarkdown": "i am working on emoji expression recognization.\ncan anyone help me?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 455119,
      "author_name": "jsaguiar",
      "author_url": "",
      "post_date": "01/13/2019 02:49:49",
      "content": "<p>Yes, just go in kernels and 'New Kernel'. You can use a notebook or script environment (in the browser) and the machine has 16gb of ram, which is probably enough for this competition.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 455145,
      "author_name": "cv13j0",
      "author_url": "",
      "post_date": "01/13/2019 04:19:18",
      "content": "<p>You can also use Google Colaboratory, with access to TPU\n<a href=\"https://colab.research.google.com/notebooks/welcome.ipynb\">https://colab.research.google.com/notebooks/welcome.ipynb</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 455539,
      "author_name": "alokpratap",
      "author_url": "",
      "post_date": "01/14/2019 06:14:33",
      "content": "<p>Yes they offer. That is enough for competition.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 456066,
      "author_name": "sriharshaatyam",
      "author_url": "",
      "post_date": "01/15/2019 03:59:11",
      "content": "<p>Yes kaggle offers enough computing resources which include 16gb ram and you can also access gpu too when you create a new kernel by clicking new kernel button in the kernels page . Also you can upload your own data in to the kaggle kernel </p>",
      "votes": null,
      "replies": [
        {
          "id": 456548,
          "author_name": "moushengxu",
          "author_url": "",
          "post_date": "01/16/2019 03:06:06",
          "content": "<p>(deleted. see below)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 456550,
      "author_name": "moushengxu",
      "author_url": "",
      "post_date": "01/16/2019 03:07:21",
      "content": "<p>Guys, thanks a lot for the info. Good to know kaggle offers the resources. The training data alone is 9.5GB. Loading them into memory, plus the N (look back time steps) * 9.5GB to feed into LSTM, I am afraid 16GB might not be enough? One has to do something smart to reduce the memory usage, I guess.</p>",
      "votes": null,
      "replies": [
        {
          "id": 456585,
          "author_name": "sriharshaatyam",
          "author_url": "",
          "post_date": "01/16/2019 05:43:19",
          "content": "<p>Hey please check this kernel to know about reducing the memory usage when loading the data into dataframe <a href=\"https://www.kaggle.com/gemartin/load-data-reduce-memory-usage\">https://www.kaggle.com/gemartin/load-data-reduce-memory-usage</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 456681,
      "author_name": "riteshand1996",
      "author_url": "",
      "post_date": "01/16/2019 09:38:44",
      "content": "<p>i am working on emoji expression recognization.\ncan anyone help me?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "455114": "Just tried to build a earthquake prediction model and soon realized that my personal computer does not have enough resources to do so. If kaggle offers computer nodes for us to upload code and run the hosted data, it would be great.\n\nThanks.",
    "455119": "Yes, just go in kernels and 'New Kernel'. You can use a notebook or script environment (in the browser) and the machine has 16gb of ram, which is probably enough for this competition.",
    "455145": "You can also use Google Colaboratory, with access to TPU\nhttps://colab.research.google.com/notebooks/welcome.ipynb",
    "455539": "Yes they offer. That is enough for competition.",
    "456066": "Yes kaggle offers enough computing resources which include 16gb ram and you can also access gpu too when you create a new kernel by clicking new kernel button in the kernels page . Also you can upload your own data in to the kaggle kernel",
    "456548": "(deleted. see below)",
    "456550": "Guys, thanks a lot for the info. Good to know kaggle offers the resources. The training data alone is 9.5GB. Loading them into memory, plus the N (look back time steps) * 9.5GB to feed into LSTM, I am afraid 16GB might not be enough? One has to do something smart to reduce the memory usage, I guess.",
    "456585": "Hey please check this kernel to know about reducing the memory usage when loading the data into dataframe https://www.kaggle.com/gemartin/load-data-reduce-memory-usage",
    "456681": "i am working on emoji expression recognization.\ncan anyone help me?"
  },
  "source": "meta"
}