{
  "id": 426698,
  "title": "Help！How does a newbie formally submit a work?",
  "url": "/competitions/asl-fingerspelling/discussion/426698",
  "author_name": "Dogge Young",
  "post_date": "2023-07-24T16:04:17.681000",
  "votes": 2,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hello, I am a junior college student new to Kaggle.  I was confused about the code structure of the submittable notebook. The introduction says to submit the model and selected columns in a zip, but the submission entry requires the notebook. So how do you make a submittable notebook? I've seen a lot of great players show their notebooks, including the model training process and zip file making. Is it necessary to include code for training models in notebooks?</p>",
  "messages": [
    {
      "id": 2357597,
      "postDate": "2023-07-25T01:51:09.347Z",
      "content": "<p>You can check, for example, this notebook for submission example: <a href=\"https://www.kaggle.com/code/gusthema/asl-fingerspelling-recognition-w-tensorflow\" target=\"_blank\">https://www.kaggle.com/code/gusthema/asl-fingerspelling-recognition-w-tensorflow</a><br>\nAlthough that notebook includes training, it is not necessary to include code for training models in notebooks.<br>\nYou can train it however you like and upload results to a Kaggle notebook.<br>\nYou'll need to use a Kaggle notebook to submit  a zip file containing two files:<br>\n!zip submission.zip  './model.tflite' './inference_args.json'<br>\nAnd then press the Submit button on the right side bar.</p>\n<p>You can start by just copying one of the example notebooks such as above, running it, and submitting.</p>\n<p>P.s: Kaggle provides free TPU computing to run notebooks. So although you don't have to train with a notebook, you may want to do so if you like to take advantage of this resource. </p>",
      "rawMarkdown": "You can check, for example, this notebook for submission example: https://www.kaggle.com/code/gusthema/asl-fingerspelling-recognition-w-tensorflow\nAlthough that notebook includes training, it is not necessary to include code for training models in notebooks.\nYou can train it however you like and upload results to a Kaggle notebook.\nYou'll need to use a Kaggle notebook to submit  a zip file containing two files:\n!zip submission.zip  './model.tflite' './inference_args.json'\nAnd then press the Submit button on the right side bar.\n\nYou can start by just copying one of the example notebooks such as above, running it, and submitting.\n\nP.s: Kaggle provides free TPU computing to run notebooks. So although you don't have to train with a notebook, you may want to do so if you like to take advantage of this resource. ",
      "votes": 3
    },
    {
      "id": 2357244,
      "postDate": "2023-07-24T16:04:17.680Z",
      "content": "<p>Hello, I am a junior college student new to Kaggle.  I was confused about the code structure of the submittable notebook. The introduction says to submit the model and selected columns in a zip, but the submission entry requires the notebook. So how do you make a submittable notebook? I've seen a lot of great players show their notebooks, including the model training process and zip file making. Is it necessary to include code for training models in notebooks?</p>",
      "rawMarkdown": "Hello, I am a junior college student new to Kaggle.  I was confused about the code structure of the submittable notebook. The introduction says to submit the model and selected columns in a zip, but the submission entry requires the notebook. So how do you make a submittable notebook? I've seen a lot of great players show their notebooks, including the model training process and zip file making. Is it necessary to include code for training models in notebooks?",
      "votes": 2
    },
    {
      "id": 2357248,
      "postDate": "2023-07-24T16:11:00.590Z",
      "content": "<p>You create your submission dataframe and export it in .csv format (or by competition rules). Once the file is present in your output folder, you can click the submit button directly in the Kaggle notebook sidebar and you are good to go.</p>",
      "rawMarkdown": "You create your submission dataframe and export it in .csv format (or by competition rules). Once the file is present in your output folder, you can click the submit button directly in the Kaggle notebook sidebar and you are good to go.",
      "votes": -3
    },
    {
      "id": 2359827,
      "postDate": "2023-07-26T11:44:02.860Z",
      "rawMarkdown": "",
      "votes": 1,
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 2357597,
      "author_name": "WalkingMoose",
      "author_url": "",
      "post_date": "2023-07-25T01:51:09.347000",
      "content": "<p>You can check, for example, this notebook for submission example: <a href=\"https://www.kaggle.com/code/gusthema/asl-fingerspelling-recognition-w-tensorflow\" target=\"_blank\">https://www.kaggle.com/code/gusthema/asl-fingerspelling-recognition-w-tensorflow</a><br>\nAlthough that notebook includes training, it is not necessary to include code for training models in notebooks.<br>\nYou can train it however you like and upload results to a Kaggle notebook.<br>\nYou'll need to use a Kaggle notebook to submit  a zip file containing two files:<br>\n!zip submission.zip  './model.tflite' './inference_args.json'<br>\nAnd then press the Submit button on the right side bar.</p>\n<p>You can start by just copying one of the example notebooks such as above, running it, and submitting.</p>\n<p>P.s: Kaggle provides free TPU computing to run notebooks. So although you don't have to train with a notebook, you may want to do so if you like to take advantage of this resource. </p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 2357248,
      "author_name": "Andrea D'Agostino",
      "author_url": "",
      "post_date": "2023-07-24T16:11:00.590000",
      "content": "<p>You create your submission dataframe and export it in .csv format (or by competition rules). Once the file is present in your output folder, you can click the submit button directly in the Kaggle notebook sidebar and you are good to go.</p>",
      "votes": -3,
      "replies": []
    },
    {
      "id": 2359827,
      "author_name": "",
      "author_url": "",
      "post_date": "2023-07-26T11:44:02.860000",
      "content": "",
      "votes": 1,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2357597": "You can check, for example, this notebook for submission example: https://www.kaggle.com/code/gusthema/asl-fingerspelling-recognition-w-tensorflow\nAlthough that notebook includes training, it is not necessary to include code for training models in notebooks.\nYou can train it however you like and upload results to a Kaggle notebook.\nYou'll need to use a Kaggle notebook to submit  a zip file containing two files:\n!zip submission.zip  './model.tflite' './inference_args.json'\nAnd then press the Submit button on the right side bar.\n\nYou can start by just copying one of the example notebooks such as above, running it, and submitting.\n\nP.s: Kaggle provides free TPU computing to run notebooks. So although you don't have to train with a notebook, you may want to do so if you like to take advantage of this resource. ",
    "2357244": "Hello, I am a junior college student new to Kaggle.  I was confused about the code structure of the submittable notebook. The introduction says to submit the model and selected columns in a zip, but the submission entry requires the notebook. So how do you make a submittable notebook? I've seen a lot of great players show their notebooks, including the model training process and zip file making. Is it necessary to include code for training models in notebooks?",
    "2357248": "You create your submission dataframe and export it in .csv format (or by competition rules). Once the file is present in your output folder, you can click the submit button directly in the Kaggle notebook sidebar and you are good to go.",
    "2359827": ""
  }
}