{
  "id": 133402,
  "title": "What is needed in Notebook submission?",
  "url": "/competitions/bengaliai-cv19/discussion/133402",
  "author_name": "",
  "post_date": "2020-03-02T15:59:17.505721300Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello, This conpetition needs notebook submissions. Then, what are notebooks required? I have to submit it with explanations for codes? or it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared? Someone see and check submission notebooks?</p>",
  "messages": [
    {
      "id": "761513",
      "postDate": "03/02/2020 15:59:17",
      "content": "<p>Hello, This conpetition needs notebook submissions. Then, what are notebooks required? I have to submit it with explanations for codes? or it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared? Someone see and check submission notebooks?</p>",
      "rawMarkdown": "Hello, This conpetition needs notebook submissions. Then, what are notebooks required? I have to submit it with explanations for codes? or it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared? Someone see and check submission notebooks?",
      "votes": null
    },
    {
      "id": "762272",
      "postDate": "03/03/2020 11:45:43",
      "content": "<p>Hi Kenta,</p>\n\n<p>Following is the response to your queries</p>\n\n<p><strong>what are notebooks required?</strong>\nNotebooks are required so that you can inference on the test images and generate <em>submission.csv</em> . The current provided test.csv contains only 12 images. Once you submit your notebook, the data is replaced with ~200,000 images. Your notebook provides inference on those images, saves them in <em>submission.csv</em>. Which is then evaluated against the ground truth and your accuracy score is generated.</p>\n\n<p><strong>I have to submit it with explanations for codes?</strong>\nNo need. All it needs to do is read the test images, run through you trained model and generate the corresponding <em>submission.csv</em></p>\n\n<p><strong>it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared?</strong>\nYes. You can do that. Provided your model is inferencing the test data and generating the <em>submission.csv</em></p>\n\n<p><strong>Someone see and check submission notebooks?</strong>\nUntil you make the notebook public. Only you can see your notebooks. No one checks the notebooks. Your result is based on the content present in <em>submission.csv</em>. That too is automatically checked.</p>",
      "rawMarkdown": "Hi Kenta,\n\nFollowing is the response to your queries\n\n**what are notebooks required?**\nNotebooks are required so that you can inference on the test images and generate _submission.csv_ . The current provided test.csv contains only 12 images. Once you submit your notebook, the data is replaced with ~200,000 images. Your notebook provides inference on those images, saves them in _submission.csv_. Which is then evaluated against the ground truth and your accuracy score is generated.\n\n**I have to submit it with explanations for codes?**\nNo need. All it needs to do is read the test images, run through you trained model and generate the corresponding _submission.csv_\n\n**it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared?**\nYes. You can do that. Provided your model is inferencing the test data and generating the _submission.csv_\n\n**Someone see and check submission notebooks?**\nUntil you make the notebook public. Only you can see your notebooks. No one checks the notebooks. Your result is based on the content present in _submission.csv_. That too is automatically checked.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 762272,
      "author_name": "santosh16k",
      "author_url": "",
      "post_date": "03/03/2020 11:45:43",
      "content": "<p>Hi Kenta,</p>\n\n<p>Following is the response to your queries</p>\n\n<p><strong>what are notebooks required?</strong>\nNotebooks are required so that you can inference on the test images and generate <em>submission.csv</em> . The current provided test.csv contains only 12 images. Once you submit your notebook, the data is replaced with ~200,000 images. Your notebook provides inference on those images, saves them in <em>submission.csv</em>. Which is then evaluated against the ground truth and your accuracy score is generated.</p>\n\n<p><strong>I have to submit it with explanations for codes?</strong>\nNo need. All it needs to do is read the test images, run through you trained model and generate the corresponding <em>submission.csv</em></p>\n\n<p><strong>it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared?</strong>\nYes. You can do that. Provided your model is inferencing the test data and generating the <em>submission.csv</em></p>\n\n<p><strong>Someone see and check submission notebooks?</strong>\nUntil you make the notebook public. Only you can see your notebooks. No one checks the notebooks. Your result is based on the content present in <em>submission.csv</em>. That too is automatically checked.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "761513": "Hello, This conpetition needs notebook submissions. Then, what are notebooks required? I have to submit it with explanations for codes? or it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared? Someone see and check submission notebooks?",
    "762272": "Hi Kenta,\n\nFollowing is the response to your queries\n\n**what are notebooks required?**\nNotebooks are required so that you can inference on the test images and generate _submission.csv_ . The current provided test.csv contains only 12 images. Once you submit your notebook, the data is replaced with ~200,000 images. Your notebook provides inference on those images, saves them in _submission.csv_. Which is then evaluated against the ground truth and your accuracy score is generated.\n\n**I have to submit it with explanations for codes?**\nNo need. All it needs to do is read the test images, run through you trained model and generate the corresponding _submission.csv_\n\n**it is OK to submit a notebook which simply loaded model from somewhere and where this model is from is undeclared?**\nYes. You can do that. Provided your model is inferencing the test data and generating the _submission.csv_\n\n**Someone see and check submission notebooks?**\nUntil you make the notebook public. Only you can see your notebooks. No one checks the notebooks. Your result is based on the content present in _submission.csv_. That too is automatically checked."
  },
  "source": "meta"
}