{
  "id": 12542,
  "title": "Faster download of dataset",
  "url": "/competitions/diabetic-retinopathy-detection/discussion/12542",
  "author_name": "MarkovC",
  "post_date": "2015-02-17T20:25:22.383000",
  "votes": 16,
  "comment_count": 44,
  "views": 16457,
  "content": "<p>To make it download faster you can do:</p>\n<p><a href=\"https://addons.mozilla.org/da/firefox/addon/downthemall/\" target=\"_blank\">DownThemAll</a> a add-on for Firefox - 5MB/s _ my max</p>\n<p>Normal Google Chrome download got 1MB/s</p>\n<p>Please share your way of making it faster so we have options.</p>",
  "messages": [
    {
      "id": 64421,
      "postDate": "2015-02-17T20:25:22.383Z",
      "content": "<p>To make it download faster you can do:</p>\n<p><a href=\"https://addons.mozilla.org/da/firefox/addon/downthemall/\" target=\"_blank\">DownThemAll</a> a add-on for Firefox - 5MB/s _ my max</p>\n<p>Normal Google Chrome download got 1MB/s</p>\n<p>Please share your way of making it faster so we have options.</p>",
      "votes": 16
    },
    {
      "id": 64524,
      "postDate": "2015-02-18T19:46:43.263Z",
      "content": "<p>As&nbsp;suggested by&nbsp;20-20 Hindsight, ec2 is a good choice. After exporting my browser cookies from firefox (with a plugin), I&nbsp;was able to download the files quite easily with wget: &quot;wget -x -c --load-cookies ~/cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/test.zip.00{1..7}&quot;</p>",
      "votes": 14
    },
    {
      "id": 64473,
      "postDate": "2015-02-18T09:00:30.920Z",
      "content": "<p>[quote=Jaco Cronje;64472]</p>\n<p>Can someone maybe center, crop and resize all the images to something like 256x256 or 512x512 and save them as raw .png files please. I really want to have a go at this competition, but the download size is just way to large to download. It will take a couple of days or weeks for me to download all the original files.</p>\n<p>I'm sure the data can be reduced to one file that is less than 8GB.</p>\n<p>[/quote]</p>\n<p>Do you also need someone to write a getting-started code with some image transformations and convolutional neural nets?</p>",
      "votes": 9
    },
    {
      "id": 64484,
      "postDate": "2015-02-18T10:40:50.833Z",
      "content": "<p>Managing the low-level plumbing is a necessary component of most Data Science projects. Have you considered creating an Amazon AWS instance and work the data there instead? If you qualify for the trial, i.e. the &quot;free tier for a year&quot;, then this data set's extra few Gb above the 30Gb free storage limit should only cost you a few cents per month.</p>\n<p>PS: At least you are not going for the<em> Microsoft Malware Classification Challenge (BIG 2015)</em> - now that one <strong>is</strong> big with around 1/2 Terabyte...&nbsp;&nbsp; :-)</p>",
      "votes": 5
    },
    {
      "id": 64475,
      "postDate": "2015-02-18T09:11:28.437Z",
      "content": "<p>Getting-started code would be nice, but I do have knowledge about that and I have my own CNN implementation, etc. I just need the data. Being located in South-Africa, it is a bit of a problem downloading so much data at reasonable speeds</p>",
      "votes": 5
    },
    {
      "id": 64553,
      "postDate": "2015-02-19T05:02:22.270Z",
      "content": "<p>[quote=20-20 Hindsight;64484]</p>\n<p>Managing the low-level plumbing is a necessary component of most Data Science projects.&nbsp;</p>\n<p>[/quote]</p>\n\n<p>While that is somewhat true we shouldn't have to all independently work out how to get around buggy zipfiles when the easiest thing to do would be to create twenty independent zip files of around 2G each where each one has the left and right eye of a patient. &nbsp;None of this creating a massive 35G zip file nonsense.</p>\n<p>Plus then people can get one of these files and start looking at images and not have to wait to get all the files off the&nbsp;slow&nbsp;kaggle2.blob.core.windows.net machine. &nbsp;Seriously it took me ~45 minutes a file on Comcast.</p>",
      "votes": 5
    },
    {
      "id": 64489,
      "postDate": "2015-02-18T13:30:20.370Z",
      "content": "<p>[quote=20-20 Hindsight;64484]</p>\n<p>Managing the low-level plumbing is a necessary component of most Data Science projects. Have you considered creating an Amazon AWS instance and work the data there instead? If you qualify for the trial, i.e. the &quot;free tier for a year&quot;, then this data set's extra few Gb above the 30Gb free storage limit should only cost you a few cents per month.</p>\n<p>PS: At least you are not going for the<em> Microsoft Malware Classification Challenge (BIG 2015)</em> - now that one <strong>is</strong> big with around 1/2 Terabyte...&nbsp;&nbsp; :-)</p>\n<p>[/quote]</p>\n<p>&lt;rant&gt;</p>\n<p>It would be great if the train set (and test set) were available in zip files that could be unzipped independently. For example, the train set files could have a certain number of files in each corresponding to a subset of patients. Having an FTP server and md5 checksums would be nice too. </p>\n<p>Think they help in management of the low level plumbing.</p>\n<p>&lt;/rant&gt;</p>",
      "votes": 3
    },
    {
      "id": 64492,
      "postDate": "2015-02-18T13:57:52.377Z",
      "content": "<p>https://www.kaggle.com/wiki/ANoteOnTorrents</p>",
      "votes": 3
    },
    {
      "id": 68281,
      "postDate": "2015-03-25T18:29:21.733Z",
      "content": "<p>Seconding @jkgiesler suggestion. The 14kb files are probably the html of our login&nbsp;page.</p>",
      "votes": 1
    },
    {
      "id": 64795,
      "postDate": "2015-02-24T00:49:32.803Z",
      "content": "<p>[quote=Jaco Cronje;64475]</p>\n<p>Getting-started code would be nice, but I do have knowledge about that and I have my own CNN implementation, etc. I just need the data. Being located in South-Africa, it is a bit of a problem downloading so much data at reasonable speeds</p>\n<p>[/quote]</p>\n<p>Hi Jaco,</p>\n<p>I had the problem with slow download speeds as well. Roughly only 1.5 Mbit/s at my home. I solve this problem by paying a little bit per month for getting a virtual server from a hosting company with a good connection (8 Euros per month for a dual core server with 6GB RAM and 500 GB storage...could of course be better in terms of processors and RAM, but it is a start). Download took me only about an hour then. The processing then is also done on the server, which can be a downside if your equipment is much better, but also an upside, because if like me, it is only a laptop, then I can use it for other work in the meantime.</p>\n<p>All the best</p>",
      "votes": 1
    },
    {
      "id": 64522,
      "postDate": "2015-02-18T18:31:41.753Z",
      "content": "<p>[quote=zeros_benchmark;64489]</p>\n<p>[quote=20-20 Hindsight;64484]</p>\n<p>Managing the low-level plumbing is a necessary component of most Data Science projects. Have you considered creating an Amazon AWS instance and work the data there instead? If you qualify for the trial, i.e. the &quot;free tier for a year&quot;, then this data set's extra few Gb above the 30Gb free storage limit should only cost you a few cents per month.</p>\n<p>PS: At least you are not going for the<em> Microsoft Malware Classification Challenge (BIG 2015)</em> - now that one <strong>is</strong> big with around 1/2 Terabyte...&nbsp;&nbsp; :-)</p>\n<p>[/quote]</p>\n<p>&lt;rant&gt;</p>\n<p>It would be great if the train set (and test set) were available in zip files that could be unzipped independently. For example, the train set files could have a certain number of files in each corresponding to a subset of patients. Having an FTP server and md5 checksums would be nice too.</p>\n<p>Think they help in management of the low level plumbing.</p>\n<p>&lt;/rant&gt;</p>\n<p>[/quote]</p>\n\n<p>I agree, the download is failing and I cannot resume.&nbsp;</p>\n<p>Please make the pieces a little smaller, say 2GB each.</p>\n<p>Thanks!</p>",
      "votes": 1
    },
    {
      "id": 64485,
      "postDate": "2015-02-18T11:28:07.010Z",
      "content": "<p>[quote=Aakash Gupta;64466]</p>\n<p>Can't these competition datasets be included as a torrent.&nbsp;</p>\n<p>I am new to these competitions, so do not know if this is feasible. But having a torrent for the data files, could be a good way to overcome a bad internet connection!</p>\n<p>[/quote]</p>\n<p>I think this topic of &quot;providing a torrent link&quot; came up before but shot down because kaggle will only want those who accept the rules to download the data and with torrents this cannot be guaranteed, apparently.</p>",
      "votes": 1
    },
    {
      "id": 68562,
      "postDate": "2015-03-27T12:16:37.887Z",
      "content": "<p>[quote=densonsmith;67322]</p>\n<p>I also tried this with the same results:</p>\n<p>wget --save-cookies cookies.txt --keep-session-cookies --post-data 'user=username&amp;password=yourpassword&#8217; http://www.kaggle.com/account/login</p>\n<p>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.00{1..5}</p>\n<p>[/quote]</p>\n<p>I think the problem here is the typo in post-data:</p>\n<p>wget --save-cookies cookies.txt --keep-session-cookies --post-data '<strong>username</strong>=yourusername&amp;password=yourpassword&#8217; http://www.kaggle.com/account/login</p>\n<p>Use 'username' instead of 'user', and I got it working.</p>",
      "votes": 2
    },
    {
      "id": 68280,
      "postDate": "2015-03-25T18:03:19.437Z",
      "content": "<p>[quote=mocany33;68274]</p>\n<p>Hi All.</p>\n<p>I have been trying to download the data set via command line using the discussion found in this competition and elsewhere using the following command</p>\n<p><strong>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.001</strong></p>\n<p>I have tried it on a number of files and I always end up with a file size of 14kbs. For instance if I were to download the&nbsp;trainLabels.csv.zip it will be a file of size 14kb instead of 64 kb.</p>\n<p>Any thoughts as to what might be going wrong.</p>\n<p>Thanks</p>\n<p>[/quote]</p>\n<p>I ran into this issue when there was an error in my cookies.txt. Double check to make sure the file contains cookies for Kaggle.com.</p>",
      "votes": 2
    },
    {
      "id": 64808,
      "postDate": "2015-02-24T15:00:16.777Z",
      "content": "<p>It's understandable. &nbsp;However, &nbsp;there's nothing stopping someone downloading the dataset and creating a torrent thereafter anyway. &nbsp;One option could be an encrypted archive, &nbsp;with the key provided after the agreement has been accepted, perhaps? &nbsp;Then again, &nbsp;we would still be reliant on x amount of people downloading the data and willing to share until the end of the competition. AWS option seems to be the best option thus far, but one&nbsp;may as well test and process there, too.</p>\n\n<p>[quote=Sashikanth Dareddy;64485]</p>\n<p>[quote=Aakash Gupta;64466]</p>\n<p>Can't these competition datasets be included as a torrent.&nbsp;</p>\n<p>I am new to these competitions, so do not know if this is feasible. But having a torrent for the data files, could be a good way to overcome a bad internet connection!</p>\n<p>[/quote]</p>\n<p>I think this topic of &quot;providing a torrent link&quot; came up before but shot down because kaggle will only want those who accept the rules to download the data and with torrents this cannot be guaranteed, apparently.</p>\n<p>[/quote]</p>",
      "votes": 2
    },
    {
      "id": 64803,
      "postDate": "2015-02-24T08:12:02.607Z",
      "content": "<p>I managed to download the full set with wget. It took a couple of days to download, but at least I have the data now. Thanks for all the tips.</p>",
      "votes": 2
    },
    {
      "id": 64466,
      "postDate": "2015-02-18T05:32:20.890Z",
      "content": "<p>Can't these competition datasets be included as a torrent.&nbsp;</p>\n<p>I am new to these competitions, so do not know if this is feasible. But having a torrent for the data files, could be a good way to overcome a bad internet connection!</p>"
    },
    {
      "id": 64472,
      "postDate": "2015-02-18T08:57:35.247Z",
      "content": "<p>Can someone maybe center, crop and resize all the images to something like 256x256 or 512x512 and save them as raw .png files please. I really want to have a go at this competition, but the download size is just way to large to download. It will take a couple of days or weeks for me to download all the original files.</p>\n<p>I'm sure the data can be reduced to one file that is less than 8GB.</p>",
      "votes": -6
    },
    {
      "id": 64885,
      "postDate": "2015-02-25T18:42:05.030Z",
      "content": "<p>I'm trying to download the training data and it's not working. please help me</p>",
      "votes": -2
    },
    {
      "id": 1220615,
      "postDate": "2021-02-28T06:59:56.077Z",
      "content": "<p>i downloaded the fundus images and after i unzip them ,produce another closed file with name fbl8ver ,,so pls how can open this file..bcz all images inside it and unreachable </p>",
      "rawMarkdown": "i downloaded the fundus images and after i unzip them ,produce another closed file with name fbl8ver ,,so pls how can open this file..bcz all images inside it and unreachable "
    },
    {
      "id": 355745,
      "postDate": "2018-07-12T09:29:50.633Z",
      "content": "<p>Hello Everyone,\nI Have one query \nWhen I am trying to download parking NYC datasets on my amazon Ec2 instance.\n1.I have followed particular instructions \nInstalling lynx browser\nTrying to login in to kaggle .com(Unable to do on my Ec2 machine facing currently)\nNow we just have to import cookies which we can do via using wget command.\nPlease help me in sorting out this issue.</p>",
      "rawMarkdown": "Hello Everyone,\nI Have one query \nWhen I am trying to download parking NYC datasets on my amazon Ec2 instance.\n1.I have followed particular instructions \nInstalling lynx browser\nTrying to login in to kaggle .com(Unable to do on my Ec2 machine facing currently)\nNow we just have to import cookies which we can do via using wget command.\nPlease help me in sorting out this issue."
    },
    {
      "id": 235100,
      "postDate": "2017-10-24T18:25:14.597Z",
      "content": "<p>I tried to use the wget option after getting the cookies.txt file, and even though i had cookies for www.kaggle.com as @jkgiesler suggested to check, it didn't work at the first trial.\nWhen i logged out of kaggle, logged in again, and got the cookies a second time - the wget worked :)\nThank you very much everyone!</p>",
      "rawMarkdown": "I tried to use the wget option after getting the cookies.txt file, and even though i had cookies for www.kaggle.com as @jkgiesler suggested to check, it didn't work at the first trial.\nWhen i logged out of kaggle, logged in again, and got the cookies a second time - the wget worked :)\nThank you very much everyone!"
    },
    {
      "id": 164789,
      "postDate": "2017-03-02T11:24:03.537Z",
      "content": "<p>open chrome or firefox developer console and switch to the network tab. Click download data and cancel the download, you'll see new records on the network activity. Right click the record and choose the \"copy as cURL\" command.\nLogin to your aws instance, paste the cURL command on your console, and add <code>-o your_filename</code> at the command. Then you can enjoy the extremely fast download on your aws instance.</p>",
      "rawMarkdown": "open chrome or firefox developer console and switch to the network tab. Click download data and cancel the download, you'll see new records on the network activity. Right click the record and choose the \"copy as cURL\" command.\nLogin to your aws instance, paste the cURL command on your console, and add `-o your_filename` at the command. Then you can enjoy the extremely fast download on your aws instance."
    },
    {
      "id": 75941,
      "postDate": "2015-05-01T16:47:19.373Z",
      "content": "<p>Aria2c does the job nicely. Just start the download in your browser, copy the ongoing download link, stop the download in the browser and paste the link in between quotes as the only parameter for Aria. If you do it in this way there is no need of cookies or authentication.</p>\n<p>Link:</p>\n<p>http://aria2.sourceforge.net/</p>"
    },
    {
      "id": 75507,
      "postDate": "2015-04-29T21:19:48.980Z",
      "content": "<p>lynx is definitely the way to go. Thank you, Joerg</p>"
    },
    {
      "id": 68590,
      "postDate": "2015-03-27T15:42:07.563Z",
      "content": "<p>Here is the Makefile script to download the dataset using curl:&nbsp;<a href=\"https://gist.github.com/nonsleepr/5f18d9b82b069948e2df\" target=\"_blank\">Kaggle Makefile</a></p>"
    },
    {
      "id": 67327,
      "postDate": "2015-03-19T22:39:03.650Z",
      "content": "<p>Maybe you should contact the administrator who responded to my frustration in the topic I referred you to earlier.&nbsp;</p>"
    },
    {
      "id": 67322,
      "postDate": "2015-03-19T22:18:39.037Z",
      "content": "<p>[quote=Sannah Ziama;67271]</p>\n<p>Hi densonsmith, I was having similar problems. I had login issues and was getting exactly the amount of bytes as you. Refer to 'error 500' topic I posted to see if that is the problem you have.</p>\n<p>In case you have same issues:</p>\n<p>&nbsp;I used wget --load-cookies just as seyn suggested but you have to do a few things first:</p>\n<p>1. make sure you can log into your Kaggle account on chrome</p>\n<p>2. get a third party plugin that handles chrome cookies</p>\n<p>3 after logging into kaggle on chrome, then download chrome cookies.txt into your directory</p>\n<p>4 wget.....</p>\n<p>[/quote]</p>\n\n<p>I also tried this with the same results:</p>\n<p>wget --save-cookies cookies.txt --keep-session-cookies --post-data 'user=username&amp;password=yourpassword&#8217; http://www.kaggle.com/account/login</p>\n\n<p>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.00{1..5}</p>"
    },
    {
      "id": 67310,
      "postDate": "2015-03-19T21:22:32.543Z",
      "content": "<p>[quote=Sannah Ziama;67271]</p>\n<p>Hi densonsmith, I was having similar problems. I had login issues and was getting exactly the amount of bytes as you. Refer to 'error 500' topic I posted to see if that is the problem you have.</p>\n<p>In case you have same issues:</p>\n<p>&nbsp;I used wget --load-cookies just as seyn suggested but you have to do a few things first:</p>\n<p>1. make sure you can log into your Kaggle account on chrome</p>\n<p>2. get a third party plugin that handles chrome cookies</p>\n<p>3 after logging into kaggle on chrome, then download chrome cookies.txt into your directory</p>\n<p>4 wget.....</p>\n<p>[/quote]</p>\n\n<p>I did all that and still got:</p>\n\n<p>Connecting to www.kaggle.com (www.kaggle.com)|168.62.224.13|:443... connected.<br>ERROR: no certificate subject alternative name matches<br> requested host name `www.kaggle.com'.<br>To connect to www.kaggle.com insecurely, use `--no-check-certificate'.</p>"
    },
    {
      "id": 67271,
      "postDate": "2015-03-19T19:14:55.893Z",
      "content": "<p>Hi densonsmith, I was having similar problems. I had login issues and was getting exactly the amount of bytes as you. Refer to 'error 500' topic I posted to see if that is the problem you have.</p>\n<p>In case you have same issues:</p>\n<p>&nbsp;I used wget --load-cookies just as seyn suggested but you have to do a few things first:</p>\n<p>1. make sure you can log into your Kaggle account on chrome</p>\n<p>2. get a third party plugin that handles chrome cookies</p>\n<p>3 after logging into kaggle on chrome, then download chrome cookies.txt into your directory</p>\n<p>4 wget.....</p>"
    },
    {
      "id": 67267,
      "postDate": "2015-03-19T18:59:46.240Z",
      "content": "<p>None of&nbsp;the suggestions about using wget are working for me. I am logged in to a remote server and I am only getting a file of 14492 bytes instead of the full file. I tried using&nbsp;--no-check-certificate but no luck. Is there some way I can log into Kaggle.com from a remote server and then download the files?</p>"
    },
    {
      "id": 66373,
      "postDate": "2015-03-16T14:47:01.660Z",
      "content": "<p>[quote=TomM;64710]</p>\n<p>I'm trying to download the training data via wget and it's not working...</p>\n<p>wget -x -c --load-cookies ~/Downloads/cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip001&nbsp;</p>\n<p>--2015-02-21 10:52:48-- https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip001<br>Resolving www.kaggle.com (www.kaggle.com)... 168.62.224.124<br>Connecting to www.kaggle.com (www.kaggle.com)|168.62.224.124|:443... connected.<br>HTTP request sent, awaiting response... 404 Not Found<br>2015-02-21 10:52:49 ERROR 404: Not Found.</p>\n<p>Any ideas what might be wrong?</p>\n<p>Never mind, it's train.zip.001 not train.zip001. It's working now.</p>\n<p>[/quote]</p>\n<p>I used &quot;wget -i files.txt&quot;, where files.txt contains the urls of &nbsp;the five training zip files, to download all training zip files and zipped them using &quot;cat train.zip.* &gt; train.zip&quot;. &nbsp;But the size of that zip file is&nbsp;72kb which I think is too small, or no? Then I tried to unzip the train.zip using kaka but got the ff:</p>\n<p>'Extraction of 'train.zip' failed</p>\n<p>Error code 2 using &quot;p7zip&quot;</p>\n<p>Fatal error'</p>\n<p>I am using OS X Yosemite.&nbsp;</p>\n<p>I also tried using p7zip on OS X 10.9.5 as suggested in other posts but I still couldn't unzip.&nbsp;</p>\n<p>Can someone offer some help?</p>"
    },
    {
      "id": 65156,
      "postDate": "2015-02-28T23:22:18.470Z",
      "content": "<p>I was just figuring that out.. pretty new to AWS. Thank you.</p>"
    },
    {
      "id": 65155,
      "postDate": "2015-02-28T23:16:36.417Z",
      "content": "<p>your root device is very small, it can't hold the whole file. You need to mount a volume and download to that.</p>"
    },
    {
      "id": 65154,
      "postDate": "2015-02-28T23:13:10.233Z",
      "content": "<p>I tried setting up an AWS instance (compute large) with 2x80 SDD and when I try and wget that file, all is well, but at about 80% of the first training set it fails and says it cant write</p>\n\n<p>```</p>\n<p>Length: 8388608000 (7.8G) [application/x-rar]<br>Saving to: &#8216;data/train.zip.001&#8217;</p>\n<p>data/train.zip.001 81%[================&gt; ] 6.38G 11.5MB/s in 6m 17s</p>\n<p><br>Cannot write to &#8216;data/train.zip.001&#8217; (Success).```</p>\n\n<p>any thoughts?</p>"
    },
    {
      "id": 65129,
      "postDate": "2015-02-28T14:09:51.093Z",
      "content": "<p>Hi Gouri, did you load your cookies with the option &quot;--load-cookies&quot; as described in my previous post?</p>"
    },
    {
      "id": 65113,
      "postDate": "2015-02-28T07:10:54.083Z",
      "content": "<p><strong>Hi All</strong></p>\n<p>Please provide me with the best possible approach to download the data. On trying with wget , I am getting connecting to Kaggle failed. Permission denied.&nbsp;</p>\n\n<p>Any inputs would be appreciated.</p>\n<p>Thanks</p>\n<p>Gouri</p>"
    },
    {
      "id": 64887,
      "postDate": "2015-02-25T19:21:09.143Z",
      "content": "<p>Could you provide some more details perhaps? Does it time out, 500 / 404 error or suchlike?</p>"
    },
    {
      "id": 64710,
      "postDate": "2015-02-21T15:55:04.467Z",
      "content": "<p>I'm trying to download the training data via wget and it's not working...</p>\n<p>wget -x -c --load-cookies ~/Downloads/cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip001&nbsp;</p>\n<p>--2015-02-21 10:52:48-- https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip001<br>Resolving www.kaggle.com (www.kaggle.com)... 168.62.224.124<br>Connecting to www.kaggle.com (www.kaggle.com)|168.62.224.124|:443... connected.<br>HTTP request sent, awaiting response... 404 Not Found<br>2015-02-21 10:52:49 ERROR 404: Not Found.</p>\n<p>Any ideas what might be wrong?</p>\n\n<p>Never mind, it's train.zip.001 not train.zip001. It's working now.</p>"
    },
    {
      "id": 64623,
      "postDate": "2015-02-20T04:11:29.177Z",
      "content": "<p>Slow internet connection here...&nbsp;</p>"
    },
    {
      "id": 64486,
      "postDate": "2015-02-18T11:29:01.623Z",
      "content": "<p>I've heard aria2 being used somewhere else on Kaggle forums.</p>"
    },
    {
      "id": 72780,
      "postDate": "2015-04-20T23:53:23.630Z",
      "content": "<p>I tried all of these suggestions to directly download into an EC2 instance and nothing worked. What finally worked for me is actually very easy: Install the lynx text-based web browser, Kaggle works nicely in it!</p>",
      "isDeleted": true
    },
    {
      "id": 69117,
      "postDate": "2015-03-31T04:17:20.867Z",
      "content": "<p>There is even better way to download datasets. I am using flareget extension in chrome on Ubuntu. You can pause and resume the downloads and it will restart where it stopped if there is problem with internet.</p>\n<p>With chrome I get a download speed of 200KB/s and flareget gives me around x10 of normal speed which is 2MB/s.</p>",
      "votes": 1,
      "isDeleted": true
    },
    {
      "id": 68284,
      "postDate": "2015-03-25T18:50:55.633Z",
      "content": "<p>Thanks&nbsp;@jkgiesler. &nbsp;Some format problem with my cookies.txt. It worked</p>",
      "isDeleted": true
    },
    {
      "id": 68274,
      "postDate": "2015-03-25T17:38:32.707Z",
      "content": "<p>Hi All.</p>\n<p>I have been trying to download the data set via command line using the discussion found in this competition and elsewhere using the following command</p>\n<p><strong>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.001</strong></p>\n<p>I have tried it on a number of files and I always end up with a file size of 14kbs. For instance if I were to download the&nbsp;trainLabels.csv.zip it will be a file of size 14kb instead of 64 kb.</p>\n<p>Any thoughts as to what might be going wrong.</p>\n\n<p>Thanks</p>",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 64524,
      "author_name": "mikael",
      "author_url": "",
      "post_date": "2015-02-18T19:46:43.263000",
      "content": "<p>As&nbsp;suggested by&nbsp;20-20 Hindsight, ec2 is a good choice. After exporting my browser cookies from firefox (with a plugin), I&nbsp;was able to download the files quite easily with wget: &quot;wget -x -c --load-cookies ~/cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/test.zip.00{1..7}&quot;</p>",
      "votes": 14,
      "replies": []
    },
    {
      "id": 64473,
      "author_name": "Abhishek Thakur",
      "author_url": "",
      "post_date": "2015-02-18T09:00:30.920000",
      "content": "<p>[quote=Jaco Cronje;64472]</p>\n<p>Can someone maybe center, crop and resize all the images to something like 256x256 or 512x512 and save them as raw .png files please. I really want to have a go at this competition, but the download size is just way to large to download. It will take a couple of days or weeks for me to download all the original files.</p>\n<p>I'm sure the data can be reduced to one file that is less than 8GB.</p>\n<p>[/quote]</p>\n<p>Do you also need someone to write a getting-started code with some image transformations and convolutional neural nets?</p>",
      "votes": 9,
      "replies": []
    },
    {
      "id": 64484,
      "author_name": "20-20 Hindsight",
      "author_url": "",
      "post_date": "2015-02-18T10:40:50.833000",
      "content": "<p>Managing the low-level plumbing is a necessary component of most Data Science projects. Have you considered creating an Amazon AWS instance and work the data there instead? If you qualify for the trial, i.e. the &quot;free tier for a year&quot;, then this data set's extra few Gb above the 30Gb free storage limit should only cost you a few cents per month.</p>\n<p>PS: At least you are not going for the<em> Microsoft Malware Classification Challenge (BIG 2015)</em> - now that one <strong>is</strong> big with around 1/2 Terabyte...&nbsp;&nbsp; :-)</p>",
      "votes": 5,
      "replies": []
    },
    {
      "id": 64475,
      "author_name": "Jaco Cronje",
      "author_url": "",
      "post_date": "2015-02-18T09:11:28.437000",
      "content": "<p>Getting-started code would be nice, but I do have knowledge about that and I have my own CNN implementation, etc. I just need the data. Being located in South-Africa, it is a bit of a problem downloading so much data at reasonable speeds</p>",
      "votes": 5,
      "replies": []
    },
    {
      "id": 64553,
      "author_name": "cwilkes",
      "author_url": "",
      "post_date": "2015-02-19T05:02:22.270000",
      "content": "<p>[quote=20-20 Hindsight;64484]</p>\n<p>Managing the low-level plumbing is a necessary component of most Data Science projects.&nbsp;</p>\n<p>[/quote]</p>\n\n<p>While that is somewhat true we shouldn't have to all independently work out how to get around buggy zipfiles when the easiest thing to do would be to create twenty independent zip files of around 2G each where each one has the left and right eye of a patient. &nbsp;None of this creating a massive 35G zip file nonsense.</p>\n<p>Plus then people can get one of these files and start looking at images and not have to wait to get all the files off the&nbsp;slow&nbsp;kaggle2.blob.core.windows.net machine. &nbsp;Seriously it took me ~45 minutes a file on Comcast.</p>",
      "votes": 5,
      "replies": []
    },
    {
      "id": 64489,
      "author_name": "zeros",
      "author_url": "",
      "post_date": "2015-02-18T13:30:20.370000",
      "content": "<p>[quote=20-20 Hindsight;64484]</p>\n<p>Managing the low-level plumbing is a necessary component of most Data Science projects. Have you considered creating an Amazon AWS instance and work the data there instead? If you qualify for the trial, i.e. the &quot;free tier for a year&quot;, then this data set's extra few Gb above the 30Gb free storage limit should only cost you a few cents per month.</p>\n<p>PS: At least you are not going for the<em> Microsoft Malware Classification Challenge (BIG 2015)</em> - now that one <strong>is</strong> big with around 1/2 Terabyte...&nbsp;&nbsp; :-)</p>\n<p>[/quote]</p>\n<p>&lt;rant&gt;</p>\n<p>It would be great if the train set (and test set) were available in zip files that could be unzipped independently. For example, the train set files could have a certain number of files in each corresponding to a subset of patients. Having an FTP server and md5 checksums would be nice too. </p>\n<p>Think they help in management of the low level plumbing.</p>\n<p>&lt;/rant&gt;</p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 64492,
      "author_name": "Will Cukierski",
      "author_url": "",
      "post_date": "2015-02-18T13:57:52.377000",
      "content": "<p>https://www.kaggle.com/wiki/ANoteOnTorrents</p>",
      "votes": 3,
      "replies": []
    },
    {
      "id": 68281,
      "author_name": "Will Cukierski",
      "author_url": "",
      "post_date": "2015-03-25T18:29:21.733000",
      "content": "<p>Seconding @jkgiesler suggestion. The 14kb files are probably the html of our login&nbsp;page.</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 64795,
      "author_name": "Kristof",
      "author_url": "",
      "post_date": "2015-02-24T00:49:32.803000",
      "content": "<p>[quote=Jaco Cronje;64475]</p>\n<p>Getting-started code would be nice, but I do have knowledge about that and I have my own CNN implementation, etc. I just need the data. Being located in South-Africa, it is a bit of a problem downloading so much data at reasonable speeds</p>\n<p>[/quote]</p>\n<p>Hi Jaco,</p>\n<p>I had the problem with slow download speeds as well. Roughly only 1.5 Mbit/s at my home. I solve this problem by paying a little bit per month for getting a virtual server from a hosting company with a good connection (8 Euros per month for a dual core server with 6GB RAM and 500 GB storage...could of course be better in terms of processors and RAM, but it is a start). Download took me only about an hour then. The processing then is also done on the server, which can be a downside if your equipment is much better, but also an upside, because if like me, it is only a laptop, then I can use it for other work in the meantime.</p>\n<p>All the best</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 64522,
      "author_name": "wagnerikeda",
      "author_url": "",
      "post_date": "2015-02-18T18:31:41.753000",
      "content": "<p>[quote=zeros_benchmark;64489]</p>\n<p>[quote=20-20 Hindsight;64484]</p>\n<p>Managing the low-level plumbing is a necessary component of most Data Science projects. Have you considered creating an Amazon AWS instance and work the data there instead? If you qualify for the trial, i.e. the &quot;free tier for a year&quot;, then this data set's extra few Gb above the 30Gb free storage limit should only cost you a few cents per month.</p>\n<p>PS: At least you are not going for the<em> Microsoft Malware Classification Challenge (BIG 2015)</em> - now that one <strong>is</strong> big with around 1/2 Terabyte...&nbsp;&nbsp; :-)</p>\n<p>[/quote]</p>\n<p>&lt;rant&gt;</p>\n<p>It would be great if the train set (and test set) were available in zip files that could be unzipped independently. For example, the train set files could have a certain number of files in each corresponding to a subset of patients. Having an FTP server and md5 checksums would be nice too.</p>\n<p>Think they help in management of the low level plumbing.</p>\n<p>&lt;/rant&gt;</p>\n<p>[/quote]</p>\n\n<p>I agree, the download is failing and I cannot resume.&nbsp;</p>\n<p>Please make the pieces a little smaller, say 2GB each.</p>\n<p>Thanks!</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 64485,
      "author_name": "Sashikanth Dareddy",
      "author_url": "",
      "post_date": "2015-02-18T11:28:07.010000",
      "content": "<p>[quote=Aakash Gupta;64466]</p>\n<p>Can't these competition datasets be included as a torrent.&nbsp;</p>\n<p>I am new to these competitions, so do not know if this is feasible. But having a torrent for the data files, could be a good way to overcome a bad internet connection!</p>\n<p>[/quote]</p>\n<p>I think this topic of &quot;providing a torrent link&quot; came up before but shot down because kaggle will only want those who accept the rules to download the data and with torrents this cannot be guaranteed, apparently.</p>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 68562,
      "author_name": "Timothy Man",
      "author_url": "",
      "post_date": "2015-03-27T12:16:37.887000",
      "content": "<p>[quote=densonsmith;67322]</p>\n<p>I also tried this with the same results:</p>\n<p>wget --save-cookies cookies.txt --keep-session-cookies --post-data 'user=username&amp;password=yourpassword&#8217; http://www.kaggle.com/account/login</p>\n<p>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.00{1..5}</p>\n<p>[/quote]</p>\n<p>I think the problem here is the typo in post-data:</p>\n<p>wget --save-cookies cookies.txt --keep-session-cookies --post-data '<strong>username</strong>=yourusername&amp;password=yourpassword&#8217; http://www.kaggle.com/account/login</p>\n<p>Use 'username' instead of 'user', and I got it working.</p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 68280,
      "author_name": "jkgiesler",
      "author_url": "",
      "post_date": "2015-03-25T18:03:19.437000",
      "content": "<p>[quote=mocany33;68274]</p>\n<p>Hi All.</p>\n<p>I have been trying to download the data set via command line using the discussion found in this competition and elsewhere using the following command</p>\n<p><strong>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.001</strong></p>\n<p>I have tried it on a number of files and I always end up with a file size of 14kbs. For instance if I were to download the&nbsp;trainLabels.csv.zip it will be a file of size 14kb instead of 64 kb.</p>\n<p>Any thoughts as to what might be going wrong.</p>\n<p>Thanks</p>\n<p>[/quote]</p>\n<p>I ran into this issue when there was an error in my cookies.txt. Double check to make sure the file contains cookies for Kaggle.com.</p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 64808,
      "author_name": "Tobi at BDSP",
      "author_url": "",
      "post_date": "2015-02-24T15:00:16.777000",
      "content": "<p>It's understandable. &nbsp;However, &nbsp;there's nothing stopping someone downloading the dataset and creating a torrent thereafter anyway. &nbsp;One option could be an encrypted archive, &nbsp;with the key provided after the agreement has been accepted, perhaps? &nbsp;Then again, &nbsp;we would still be reliant on x amount of people downloading the data and willing to share until the end of the competition. AWS option seems to be the best option thus far, but one&nbsp;may as well test and process there, too.</p>\n\n<p>[quote=Sashikanth Dareddy;64485]</p>\n<p>[quote=Aakash Gupta;64466]</p>\n<p>Can't these competition datasets be included as a torrent.&nbsp;</p>\n<p>I am new to these competitions, so do not know if this is feasible. But having a torrent for the data files, could be a good way to overcome a bad internet connection!</p>\n<p>[/quote]</p>\n<p>I think this topic of &quot;providing a torrent link&quot; came up before but shot down because kaggle will only want those who accept the rules to download the data and with torrents this cannot be guaranteed, apparently.</p>\n<p>[/quote]</p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 64803,
      "author_name": "Jaco Cronje",
      "author_url": "",
      "post_date": "2015-02-24T08:12:02.607000",
      "content": "<p>I managed to download the full set with wget. It took a couple of days to download, but at least I have the data now. Thanks for all the tips.</p>",
      "votes": 2,
      "replies": []
    },
    {
      "id": 64466,
      "author_name": "SkyLord",
      "author_url": "",
      "post_date": "2015-02-18T05:32:20.890000",
      "content": "<p>Can't these competition datasets be included as a torrent.&nbsp;</p>\n<p>I am new to these competitions, so do not know if this is feasible. But having a torrent for the data files, could be a good way to overcome a bad internet connection!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 64472,
      "author_name": "Jaco Cronje",
      "author_url": "",
      "post_date": "2015-02-18T08:57:35.247000",
      "content": "<p>Can someone maybe center, crop and resize all the images to something like 256x256 or 512x512 and save them as raw .png files please. I really want to have a go at this competition, but the download size is just way to large to download. It will take a couple of days or weeks for me to download all the original files.</p>\n<p>I'm sure the data can be reduced to one file that is less than 8GB.</p>",
      "votes": -6,
      "replies": []
    },
    {
      "id": 64885,
      "author_name": "lisa ahmed",
      "author_url": "",
      "post_date": "2015-02-25T18:42:05.030000",
      "content": "<p>I'm trying to download the training data and it's not working. please help me</p>",
      "votes": -2,
      "replies": []
    },
    {
      "id": 1220615,
      "author_name": "smsinan",
      "author_url": "",
      "post_date": "2021-02-28T06:59:56.077000",
      "content": "<p>i downloaded the fundus images and after i unzip them ,produce another closed file with name fbl8ver ,,so pls how can open this file..bcz all images inside it and unreachable </p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 355745,
      "author_name": "Sachin Rungta",
      "author_url": "",
      "post_date": "2018-07-12T09:29:50.633000",
      "content": "<p>Hello Everyone,\nI Have one query \nWhen I am trying to download parking NYC datasets on my amazon Ec2 instance.\n1.I have followed particular instructions \nInstalling lynx browser\nTrying to login in to kaggle .com(Unable to do on my Ec2 machine facing currently)\nNow we just have to import cookies which we can do via using wget command.\nPlease help me in sorting out this issue.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 235100,
      "author_name": "Gal Avineri",
      "author_url": "",
      "post_date": "2017-10-24T18:25:14.597000",
      "content": "<p>I tried to use the wget option after getting the cookies.txt file, and even though i had cookies for www.kaggle.com as @jkgiesler suggested to check, it didn't work at the first trial.\nWhen i logged out of kaggle, logged in again, and got the cookies a second time - the wget worked :)\nThank you very much everyone!</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 164789,
      "author_name": "Chih-Cheng Liang",
      "author_url": "",
      "post_date": "2017-03-02T11:24:03.537000",
      "content": "<p>open chrome or firefox developer console and switch to the network tab. Click download data and cancel the download, you'll see new records on the network activity. Right click the record and choose the \"copy as cURL\" command.\nLogin to your aws instance, paste the cURL command on your console, and add <code>-o your_filename</code> at the command. Then you can enjoy the extremely fast download on your aws instance.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 75941,
      "author_name": "Alejandro Mosquera",
      "author_url": "",
      "post_date": "2015-05-01T16:47:19.373000",
      "content": "<p>Aria2c does the job nicely. Just start the download in your browser, copy the ongoing download link, stop the download in the browser and paste the link in between quotes as the only parameter for Aria. If you do it in this way there is no need of cookies or authentication.</p>\n<p>Link:</p>\n<p>http://aria2.sourceforge.net/</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 75507,
      "author_name": "Cole Diamond",
      "author_url": "",
      "post_date": "2015-04-29T21:19:48.980000",
      "content": "<p>lynx is definitely the way to go. Thank you, Joerg</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 68590,
      "author_name": "nonsleepr",
      "author_url": "",
      "post_date": "2015-03-27T15:42:07.563000",
      "content": "<p>Here is the Makefile script to download the dataset using curl:&nbsp;<a href=\"https://gist.github.com/nonsleepr/5f18d9b82b069948e2df\" target=\"_blank\">Kaggle Makefile</a></p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 67327,
      "author_name": "Sannah Ziama",
      "author_url": "",
      "post_date": "2015-03-19T22:39:03.650000",
      "content": "<p>Maybe you should contact the administrator who responded to my frustration in the topic I referred you to earlier.&nbsp;</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 67322,
      "author_name": "densonsmith",
      "author_url": "",
      "post_date": "2015-03-19T22:18:39.037000",
      "content": "<p>[quote=Sannah Ziama;67271]</p>\n<p>Hi densonsmith, I was having similar problems. I had login issues and was getting exactly the amount of bytes as you. Refer to 'error 500' topic I posted to see if that is the problem you have.</p>\n<p>In case you have same issues:</p>\n<p>&nbsp;I used wget --load-cookies just as seyn suggested but you have to do a few things first:</p>\n<p>1. make sure you can log into your Kaggle account on chrome</p>\n<p>2. get a third party plugin that handles chrome cookies</p>\n<p>3 after logging into kaggle on chrome, then download chrome cookies.txt into your directory</p>\n<p>4 wget.....</p>\n<p>[/quote]</p>\n\n<p>I also tried this with the same results:</p>\n<p>wget --save-cookies cookies.txt --keep-session-cookies --post-data 'user=username&amp;password=yourpassword&#8217; http://www.kaggle.com/account/login</p>\n\n<p>wget -x -c --load-cookies cookies.txt -P data -nH --cut-dirs=5 https://www.kaggle.com/c/diabetic-retinopathy-detection/download/train.zip.00{1..5}</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 67310,
      "author_name": "densonsmith",
      "author_url": "",
      "post_date": "2015-03-19T21:22:32.543000",
      "content": "<p>[quote=Sannah Ziama;67271]</p>\n<p>Hi densonsmith, I was having similar problems. I had login issues and was getting exactly the amount of bytes as you. Refer to 'error 500' topic I posted to see if that is the problem you have.</p>\n<p>In case you have same issues:</p>\n<p>&nbsp;I used wget --load-cookies just as seyn suggested but you have to do a few things first:</p>\n<p>1. make sure you can log into your Kaggle account on chrome</p>\n<p>2. get a third party plugin that handles chrome cookies</p>\n<p>3 after logging into kaggle on chrome, then download chrome cookies.txt into your directory</p>\n<p>4 wget.....</p>\n<p>[/quote]</p>\n\n<p>I did all that and still got:</p>\n\n<p>Connecting to www.kaggle.com (www.kaggle.com)|168.62.224.13|:443... connected.<br>ERROR: no certificate subject alternative name matches<br> requested host name `www.kaggle.com'.<br>To connect to www.kaggle.com insecurely, use `--no-check-certificate'.</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 67271,
      "author_name": "Sannah Ziama",
      "author_url": "",
      "post_date": "2015-03-19T19:14:55.893000",
      "content": "<p>Hi densonsmith, I was having similar problems. I had login issues and was getting exactly the amount of bytes as you. Refer to 'error 500' topic I posted to see if that is the problem you have.</p>\n<p>In case you have same issues:</p>\n<p>&nbsp;I used wget --load-cookies just as seyn suggested but you have to do a few things first:</p>\n<p>1. make sure you can log into your Kaggle account on chrome</p>\n<p>2. get a third party plugin that handles chrome cookies</p>\n<p>3 after logging into kaggle on chrome, then download chrome cookies.txt into your directory</p>\n<p>4 wget.....</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 67267,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-03-19T18:59:46.240000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 66373,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-03-16T14:47:01.660000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 65156,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-28T23:22:18.470000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 65155,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-28T23:16:36.417000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 65154,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-28T23:13:10.233000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 65129,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-28T14:09:51.093000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 65113,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-28T07:10:54.083000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 64887,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-25T19:21:09.143000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 64710,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-21T15:55:04.467000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 64623,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-20T04:11:29.177000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 64486,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-02-18T11:29:01.623000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 72780,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-04-20T23:53:23.630000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 69117,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-03-31T04:17:20.867000",
      "content": "",
      "votes": 1,
      "replies": []
    },
    {
      "id": 68284,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-03-25T18:50:55.633000",
      "content": "",
      "votes": 0,
      "replies": []
    },
    {
      "id": 68274,
      "author_name": "",
      "author_url": "",
      "post_date": "2015-03-25T17:38:32.707000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "64421": "",
    "64524": "",
    "64473": "",
    "64484": "",
    "64475": "",
    "64553": "",
    "64489": "",
    "64492": "",
    "68281": "",
    "64795": "",
    "64522": "",
    "64485": "",
    "68562": "",
    "68280": "",
    "64808": "",
    "64803": "",
    "64466": "",
    "64472": "",
    "64885": "",
    "1220615": "i downloaded the fundus images and after i unzip them ,produce another closed file with name fbl8ver ,,so pls how can open this file..bcz all images inside it and unreachable ",
    "355745": "Hello Everyone,\nI Have one query \nWhen I am trying to download parking NYC datasets on my amazon Ec2 instance.\n1.I have followed particular instructions \nInstalling lynx browser\nTrying to login in to kaggle .com(Unable to do on my Ec2 machine facing currently)\nNow we just have to import cookies which we can do via using wget command.\nPlease help me in sorting out this issue.",
    "235100": "I tried to use the wget option after getting the cookies.txt file, and even though i had cookies for www.kaggle.com as @jkgiesler suggested to check, it didn't work at the first trial.\nWhen i logged out of kaggle, logged in again, and got the cookies a second time - the wget worked :)\nThank you very much everyone!",
    "164789": "open chrome or firefox developer console and switch to the network tab. Click download data and cancel the download, you'll see new records on the network activity. Right click the record and choose the \"copy as cURL\" command.\nLogin to your aws instance, paste the cURL command on your console, and add `-o your_filename` at the command. Then you can enjoy the extremely fast download on your aws instance.",
    "75941": "",
    "75507": "",
    "68590": "",
    "67327": "",
    "67322": "",
    "67310": "",
    "67271": "",
    "67267": "",
    "66373": "",
    "65156": "",
    "65155": "",
    "65154": "",
    "65129": "",
    "65113": "",
    "64887": "",
    "64710": "",
    "64623": "",
    "64486": "",
    "72780": "",
    "69117": "",
    "68284": "",
    "68274": ""
  }
}