{
  "id": 98807,
  "title": "Please, Help me resolve the error ",
  "url": "/competitions/aptos2019-blindness-detection/discussion/98807",
  "author_name": "",
  "post_date": "2019-07-06T14:36:55.534178300Z",
  "votes": 3,
  "comment_count": 8,
  "views": 0,
  "content": "<p><a href=\"https://www.kaggle.com/iluvmahheart/simple-beginner-blindness-detector\">https://www.kaggle.com/iluvmahheart/simple-beginner-blindness-detector</a></p>",
  "messages": [
    {
      "id": "569368",
      "postDate": "07/06/2019 14:36:55",
      "content": "<p><a href=\"https://www.kaggle.com/iluvmahheart/simple-beginner-blindness-detector\">https://www.kaggle.com/iluvmahheart/simple-beginner-blindness-detector</a></p>",
      "rawMarkdown": "https://www.kaggle.com/iluvmahheart/simple-beginner-blindness-detector",
      "votes": null
    },
    {
      "id": "569373",
      "postDate": "07/06/2019 14:40:52",
      "content": "<p>Try Turning it on and off again</p>",
      "rawMarkdown": "Try Turning it on and off again",
      "votes": null
    },
    {
      "id": "569575",
      "postDate": "07/07/2019 00:51:30",
      "content": "<p>which error? i don't find any error there!!!</p>",
      "rawMarkdown": "which error? i don't find any error there!!!",
      "votes": null
    },
    {
      "id": "569708",
      "postDate": "07/07/2019 07:31:40",
      "content": "<p>This model has activation='softmax' and this Dense's classes =5 .\nI think preds is has 5 probabilities.\nSo you should preds that the index has max probabilities.\nI am sorry for my unskillful English. </p>",
      "rawMarkdown": "This model has activation='softmax' and this Dense's classes =5 .\nI think preds is has 5 probabilities.\nSo you should preds that the index has max probabilities.\nI am sorry for my unskillful English.",
      "votes": null
    },
    {
      "id": "569720",
      "postDate": "07/07/2019 07:45:58",
      "content": "<p><a href=\"/mobassir\">@mobassir</a>  i want prediction output as 0,1,2,3,4 but i am getting float outputs , so my output is not matching sample output </p>",
      "rawMarkdown": "mobassir  i want prediction output as 0,1,2,3,4 but i am getting float outputs , so my output is not matching sample output",
      "votes": null
    },
    {
      "id": "569721",
      "postDate": "07/07/2019 07:46:53",
      "content": "<p>yes, i am unable to understand please clarify :(</p>",
      "rawMarkdown": "yes, i am unable to understand please clarify :(",
      "votes": null
    },
    {
      "id": "569726",
      "postDate": "07/07/2019 08:04:42",
      "content": "<p>Maybe you should try following.\nin[27]\n<code>\npreds = model.predict(test_X) \n</code>\n<code>\npreds_max = np.argmax(preds,axis = 1) <br>\n</code>\n<code>\ndf_submission['diagnosis']= preds_max\n</code></p>\n\n<p><a href=\"https://www.kaggle.com/snakayama/vgg16-keras?scriptVersionId=16843184\">https://www.kaggle.com/snakayama/vgg16-keras?scriptVersionId=16843184</a></p>",
      "rawMarkdown": "Maybe you should try following.\nin[27]\n`\npreds = model.predict(test_X) \n`\n`\npreds_max = np.argmax(preds,axis = 1)  \n`\n` \ndf_submission['diagnosis']= preds_max\n`\n\nhttps://www.kaggle.com/snakayama/vgg16-keras?scriptVersionId=16843184",
      "votes": null
    },
    {
      "id": "569760",
      "postDate": "07/07/2019 09:15:41",
      "content": "<p><a href=\"/snakayama\">@snakayama</a> Thanks a lot it worked 😻 </p>",
      "rawMarkdown": "snakayama Thanks a lot it worked 😻",
      "votes": null
    },
    {
      "id": "569841",
      "postDate": "07/07/2019 12:38:29",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2034058%2F64ec3363eb0f9c0ebc5a8785078f2e8a%2FoneHot.png?generation=1562503024824584&amp;alt=media\" alt=\"\">\ni can clearly see your predicted outputs are in one hot encoded format and not in floating point format</p>",
      "rawMarkdown": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2034058%2F64ec3363eb0f9c0ebc5a8785078f2e8a%2FoneHot.png?generation=1562503024824584&amp;alt=media)\ni can clearly see your predicted outputs are in one hot encoded format and not in floating point format",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 569373,
      "author_name": "ishmeet",
      "author_url": "",
      "post_date": "07/06/2019 14:40:52",
      "content": "<p>Try Turning it on and off again</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 569575,
      "author_name": "mobassir",
      "author_url": "",
      "post_date": "07/07/2019 00:51:30",
      "content": "<p>which error? i don't find any error there!!!</p>",
      "votes": null,
      "replies": [
        {
          "id": 569720,
          "author_name": "iluvmahheart",
          "author_url": "",
          "post_date": "07/07/2019 07:45:58",
          "content": "<p><a href=\"/mobassir\">@mobassir</a>  i want prediction output as 0,1,2,3,4 but i am getting float outputs , so my output is not matching sample output </p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 569841,
          "author_name": "mobassir",
          "author_url": "",
          "post_date": "07/07/2019 12:38:29",
          "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2034058%2F64ec3363eb0f9c0ebc5a8785078f2e8a%2FoneHot.png?generation=1562503024824584&amp;alt=media\" alt=\"\">\ni can clearly see your predicted outputs are in one hot encoded format and not in floating point format</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 569708,
      "author_name": "snakayama",
      "author_url": "",
      "post_date": "07/07/2019 07:31:40",
      "content": "<p>This model has activation='softmax' and this Dense's classes =5 .\nI think preds is has 5 probabilities.\nSo you should preds that the index has max probabilities.\nI am sorry for my unskillful English. </p>",
      "votes": null,
      "replies": [
        {
          "id": 569721,
          "author_name": "iluvmahheart",
          "author_url": "",
          "post_date": "07/07/2019 07:46:53",
          "content": "<p>yes, i am unable to understand please clarify :(</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 569726,
          "author_name": "snakayama",
          "author_url": "",
          "post_date": "07/07/2019 08:04:42",
          "content": "<p>Maybe you should try following.\nin[27]\n<code>\npreds = model.predict(test_X) \n</code>\n<code>\npreds_max = np.argmax(preds,axis = 1) <br>\n</code>\n<code>\ndf_submission['diagnosis']= preds_max\n</code></p>\n\n<p><a href=\"https://www.kaggle.com/snakayama/vgg16-keras?scriptVersionId=16843184\">https://www.kaggle.com/snakayama/vgg16-keras?scriptVersionId=16843184</a></p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 569760,
          "author_name": "iluvmahheart",
          "author_url": "",
          "post_date": "07/07/2019 09:15:41",
          "content": "<p><a href=\"/snakayama\">@snakayama</a> Thanks a lot it worked 😻 </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "569368": "https://www.kaggle.com/iluvmahheart/simple-beginner-blindness-detector",
    "569373": "Try Turning it on and off again",
    "569575": "which error? i don't find any error there!!!",
    "569708": "This model has activation='softmax' and this Dense's classes =5 .\nI think preds is has 5 probabilities.\nSo you should preds that the index has max probabilities.\nI am sorry for my unskillful English.",
    "569720": "mobassir  i want prediction output as 0,1,2,3,4 but i am getting float outputs , so my output is not matching sample output",
    "569721": "yes, i am unable to understand please clarify :(",
    "569726": "Maybe you should try following.\nin[27]\n`\npreds = model.predict(test_X) \n`\n`\npreds_max = np.argmax(preds,axis = 1)  \n`\n` \ndf_submission['diagnosis']= preds_max\n`\n\nhttps://www.kaggle.com/snakayama/vgg16-keras?scriptVersionId=16843184",
    "569760": "snakayama Thanks a lot it worked 😻",
    "569841": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2034058%2F64ec3363eb0f9c0ebc5a8785078f2e8a%2FoneHot.png?generation=1562503024824584&amp;alt=media)\ni can clearly see your predicted outputs are in one hot encoded format and not in floating point format"
  },
  "source": "meta"
}