{
  "id": 478206,
  "title": "Is possible to use a notebook to process a database, save the output and load in another notebook to model?",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/478206",
  "author_name": "",
  "post_date": "2024-02-19T16:53:52.675414100Z",
  "votes": 3,
  "comment_count": 5,
  "views": 0,
  "content": "<p>I'm new to competitions, is possible to use a notebook to process data and another to model da result and predict the output?</p>",
  "messages": [
    {
      "id": "2659190",
      "postDate": "02/19/2024 16:53:52",
      "content": "<p>I'm new to competitions, is possible to use a notebook to process data and another to model da result and predict the output?</p>",
      "rawMarkdown": "I'm new to competitions, is possible to use a notebook to process data and another to model da result and predict the output?",
      "votes": null
    },
    {
      "id": "2659203",
      "postDate": "02/19/2024 17:01:01",
      "content": "<p>Yes, you need to attach the first notebook’s output as a data source to the second one or save it as a dataset, then attach that dataset to the second notebook. You may get some insights by going through <a href=\"https://www.kaggle.com/code/kononenko/pip-install-no-internet\" target=\"_blank\">pip-install-no-internet</a> kernel, just instead of the python wheels it is going to be your custom data.</p>",
      "rawMarkdown": "Yes, you need to attach the first notebook’s output as a data source to the second one or save it as a dataset, then attach that dataset to the second notebook. You may get some insights by going through [pip-install-no-internet](https://www.kaggle.com/code/kononenko/pip-install-no-internet) kernel, just instead of the python wheels it is going to be your custom data.",
      "votes": null
    },
    {
      "id": "2659658",
      "postDate": "02/20/2024 04:37:10",
      "content": "<p>Yes of course, this is a preferred way for code comps <a href=\"https://www.kaggle.com/alexandreg1998\" target=\"_blank\">@alexandreg1998</a> </p>",
      "rawMarkdown": "Yes of course, this is a preferred way for code comps @alexandreg1998",
      "votes": null
    },
    {
      "id": "2768195",
      "postDate": "04/22/2024 18:11:32",
      "content": "<p>This works for me with lightGBM. I can't get my neural network model, saved as .keras to open as a data set or as a model. Do you know what I am missing?</p>\n<p>I've uploaded many lightGBM models without problems. I think it is something about .keras </p>\n<p>From this code<br>\nbest_model = tf.keras.models.load_model('/kaggle/input/nndataset/best_model.keras')<br>\nbest_model = tf.keras.models.load_model('/kaggle/input/nn-model/keras/nn/1/best_model.keras')</p>\n<p>I get the error: OSError: [Errno 30] Read-only file system: '/kaggle/input/nndataset/best_model.keras'</p>",
      "rawMarkdown": "This works for me with lightGBM. I can't get my neural network model, saved as .keras to open as a data set or as a model. Do you know what I am missing?\n\nI've uploaded many lightGBM models without problems. I think it is something about .keras \n\nFrom this code\nbest_model = tf.keras.models.load_model('/kaggle/input/nndataset/best_model.keras')\nbest_model = tf.keras.models.load_model('/kaggle/input/nn-model/keras/nn/1/best_model.keras')\n\nI get the error: OSError: [Errno 30] Read-only file system: '/kaggle/input/nndataset/best_model.keras'",
      "votes": null
    },
    {
      "id": "2768474",
      "postDate": "04/22/2024 21:22:56",
      "content": "<p><a href=\"https://www.kaggle.com/nicksalem\" target=\"_blank\">@nicksalem</a> Unfortunately, this is the first time I see such an issue and it doesn't seem to be related to what you're doing. I recommend you to contact Kaggle support on this matter.</p>",
      "rawMarkdown": "nicksalem Unfortunately, this is the first time I see such an issue and it doesn't seem to be related to what you're doing. I recommend you to contact Kaggle support on this matter.",
      "votes": null
    },
    {
      "id": "2768479",
      "postDate": "04/22/2024 21:27:08",
      "content": "<p>Thanks for the response :)</p>",
      "rawMarkdown": "Thanks for the response :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2659203,
      "author_name": "kononenko",
      "author_url": "",
      "post_date": "02/19/2024 17:01:01",
      "content": "<p>Yes, you need to attach the first notebook’s output as a data source to the second one or save it as a dataset, then attach that dataset to the second notebook. You may get some insights by going through <a href=\"https://www.kaggle.com/code/kononenko/pip-install-no-internet\" target=\"_blank\">pip-install-no-internet</a> kernel, just instead of the python wheels it is going to be your custom data.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2768195,
          "author_name": "nicksalem",
          "author_url": "",
          "post_date": "04/22/2024 18:11:32",
          "content": "<p>This works for me with lightGBM. I can't get my neural network model, saved as .keras to open as a data set or as a model. Do you know what I am missing?</p>\n<p>I've uploaded many lightGBM models without problems. I think it is something about .keras </p>\n<p>From this code<br>\nbest_model = tf.keras.models.load_model('/kaggle/input/nndataset/best_model.keras')<br>\nbest_model = tf.keras.models.load_model('/kaggle/input/nn-model/keras/nn/1/best_model.keras')</p>\n<p>I get the error: OSError: [Errno 30] Read-only file system: '/kaggle/input/nndataset/best_model.keras'</p>",
          "votes": null,
          "replies": [
            {
              "id": 2768474,
              "author_name": "kononenko",
              "author_url": "",
              "post_date": "04/22/2024 21:22:56",
              "content": "<p><a href=\"https://www.kaggle.com/nicksalem\" target=\"_blank\">@nicksalem</a> Unfortunately, this is the first time I see such an issue and it doesn't seem to be related to what you're doing. I recommend you to contact Kaggle support on this matter.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 2768479,
                  "author_name": "nicksalem",
                  "author_url": "",
                  "post_date": "04/22/2024 21:27:08",
                  "content": "<p>Thanks for the response :)</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    },
    {
      "id": 2659658,
      "author_name": "ravi20076",
      "author_url": "",
      "post_date": "02/20/2024 04:37:10",
      "content": "<p>Yes of course, this is a preferred way for code comps <a href=\"https://www.kaggle.com/alexandreg1998\" target=\"_blank\">@alexandreg1998</a> </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2659190": "I'm new to competitions, is possible to use a notebook to process data and another to model da result and predict the output?",
    "2659203": "Yes, you need to attach the first notebook’s output as a data source to the second one or save it as a dataset, then attach that dataset to the second notebook. You may get some insights by going through [pip-install-no-internet](https://www.kaggle.com/code/kononenko/pip-install-no-internet) kernel, just instead of the python wheels it is going to be your custom data.",
    "2659658": "Yes of course, this is a preferred way for code comps @alexandreg1998",
    "2768195": "This works for me with lightGBM. I can't get my neural network model, saved as .keras to open as a data set or as a model. Do you know what I am missing?\n\nI've uploaded many lightGBM models without problems. I think it is something about .keras \n\nFrom this code\nbest_model = tf.keras.models.load_model('/kaggle/input/nndataset/best_model.keras')\nbest_model = tf.keras.models.load_model('/kaggle/input/nn-model/keras/nn/1/best_model.keras')\n\nI get the error: OSError: [Errno 30] Read-only file system: '/kaggle/input/nndataset/best_model.keras'",
    "2768474": "nicksalem Unfortunately, this is the first time I see such an issue and it doesn't seem to be related to what you're doing. I recommend you to contact Kaggle support on this matter.",
    "2768479": "Thanks for the response :)"
  },
  "source": "meta"
}