{
  "id": 174174,
  "title": "KaggleDatasets().get_gcs_path() on colab",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/174174",
  "author_name": "",
  "post_date": "2020-08-12T14:42:04.476278800Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>My TPU quota is completed on kaggle and I switched to Colab. I download the TF records dataset using my API token but how to get the gcs path on colab. It is throwing ModuleNotFoundError on importing  <code>from kaggle_datasets import KaggleDatasets</code>. Is there an alternate way to use TFrecords on colab?</p>",
  "messages": [
    {
      "id": "967856",
      "postDate": "08/12/2020 14:42:04",
      "content": "<p>My TPU quota is completed on kaggle and I switched to Colab. I download the TF records dataset using my API token but how to get the gcs path on colab. It is throwing ModuleNotFoundError on importing  <code>from kaggle_datasets import KaggleDatasets</code>. Is there an alternate way to use TFrecords on colab?</p>",
      "rawMarkdown": "My TPU quota is completed on kaggle and I switched to Colab. I download the TF records dataset using my API token but how to get the gcs path on colab. It is throwing ModuleNotFoundError on importing  `from kaggle_datasets import KaggleDatasets`. Is there an alternate way to use TFrecords on colab?",
      "votes": null
    },
    {
      "id": "967897",
      "postDate": "08/12/2020 15:13:00",
      "content": "<p>You don't have to download the dataset. Find the GCS path of the dataset you are using from a Kaggle notebook and you can directly use the path in the colab notebook. The path would be something like this.</p>\n<p>GCS_PATH = 'gs://kds-71b679af6661316328cfa122f5342b4a84fe0c7c1135d809fbee8ca4'</p>",
      "rawMarkdown": "You don't have to download the dataset. Find the GCS path of the dataset you are using from a Kaggle notebook and you can directly use the path in the colab notebook. The path would be something like this.\n\n GCS_PATH = 'gs://kds-71b679af6661316328cfa122f5342b4a84fe0c7c1135d809fbee8ca4'",
      "votes": null
    },
    {
      "id": "967941",
      "postDate": "08/12/2020 15:40:42",
      "content": "<p>Thank you Jose!</p>",
      "rawMarkdown": "Thank you Jose!",
      "votes": null
    },
    {
      "id": "968860",
      "postDate": "08/13/2020 09:36:57",
      "content": "<p>Hi,<br>\nyet another solution (instead of copy paste) I've developed and described here:<br>\n <a href=\"https://www.kaggle.com/wrrosa/gcs-bucket-addresses-utility-package\" target=\"_blank\">https://www.kaggle.com/wrrosa/gcs-bucket-addresses-utility-package</a></p>",
      "rawMarkdown": "Hi,\nyet another solution (instead of copy paste) I've developed and described here:\n https://www.kaggle.com/wrrosa/gcs-bucket-addresses-utility-package",
      "votes": null
    },
    {
      "id": "968861",
      "postDate": "08/13/2020 09:38:59",
      "content": "<p><a href=\"https://www.kaggle.com/aakashveera\" target=\"_blank\">@aakashveera</a> <br>\nSee this -&gt; <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/174299\" target=\"_blank\">https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/174299</a><br>\nscroll down then you will see my code<br>\nsorry for not being accustomed to MARKDOWN</p>",
      "rawMarkdown": "aakashveera \nSee this -> https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/174299\nscroll down then you will see my code\nsorry for not being accustomed to MARKDOWN",
      "votes": null
    },
    {
      "id": "1280946",
      "postDate": "04/22/2021 13:56:28",
      "content": "<p>Thanks Jose.</p>",
      "rawMarkdown": "Thanks Jose.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 967897,
      "author_name": "josealways123",
      "author_url": "",
      "post_date": "08/12/2020 15:13:00",
      "content": "<p>You don't have to download the dataset. Find the GCS path of the dataset you are using from a Kaggle notebook and you can directly use the path in the colab notebook. The path would be something like this.</p>\n<p>GCS_PATH = 'gs://kds-71b679af6661316328cfa122f5342b4a84fe0c7c1135d809fbee8ca4'</p>",
      "votes": null,
      "replies": [
        {
          "id": 967941,
          "author_name": "aakashveera",
          "author_url": "",
          "post_date": "08/12/2020 15:40:42",
          "content": "<p>Thank you Jose!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1280946,
          "author_name": "aifahim",
          "author_url": "",
          "post_date": "04/22/2021 13:56:28",
          "content": "<p>Thanks Jose.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 968860,
      "author_name": "wrrosa",
      "author_url": "",
      "post_date": "08/13/2020 09:36:57",
      "content": "<p>Hi,<br>\nyet another solution (instead of copy paste) I've developed and described here:<br>\n <a href=\"https://www.kaggle.com/wrrosa/gcs-bucket-addresses-utility-package\" target=\"_blank\">https://www.kaggle.com/wrrosa/gcs-bucket-addresses-utility-package</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 968861,
      "author_name": "deepkim",
      "author_url": "",
      "post_date": "08/13/2020 09:38:59",
      "content": "<p><a href=\"https://www.kaggle.com/aakashveera\" target=\"_blank\">@aakashveera</a> <br>\nSee this -&gt; <a href=\"https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/174299\" target=\"_blank\">https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/174299</a><br>\nscroll down then you will see my code<br>\nsorry for not being accustomed to MARKDOWN</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "967856": "My TPU quota is completed on kaggle and I switched to Colab. I download the TF records dataset using my API token but how to get the gcs path on colab. It is throwing ModuleNotFoundError on importing  `from kaggle_datasets import KaggleDatasets`. Is there an alternate way to use TFrecords on colab?",
    "967897": "You don't have to download the dataset. Find the GCS path of the dataset you are using from a Kaggle notebook and you can directly use the path in the colab notebook. The path would be something like this.\n\n GCS_PATH = 'gs://kds-71b679af6661316328cfa122f5342b4a84fe0c7c1135d809fbee8ca4'",
    "967941": "Thank you Jose!",
    "968860": "Hi,\nyet another solution (instead of copy paste) I've developed and described here:\n https://www.kaggle.com/wrrosa/gcs-bucket-addresses-utility-package",
    "968861": "aakashveera \nSee this -> https://www.kaggle.com/c/siim-isic-melanoma-classification/discussion/174299\nscroll down then you will see my code\nsorry for not being accustomed to MARKDOWN",
    "1280946": "Thanks Jose."
  },
  "source": "meta"
}