{
  "total": 217,
  "topics": [
    {
      "id": 25428,
      "title": "Official pre-trained model and data thread",
      "comment_count": 189,
      "views": 16945,
      "votes": 9
    },
    {
      "id": 32153,
      "title": "Congrats to the winners",
      "comment_count": 18,
      "views": 2,
      "votes": 14
    },
    {
      "id": 31624,
      "title": "Multiple account holders (cheaters) removed",
      "comment_count": 4,
      "views": 0,
      "votes": 8
    },
    {
      "id": 30805,
      "title": "Details about second stage test data",
      "comment_count": 103,
      "views": 0,
      "votes": 19
    },
    {
      "id": 31103,
      "title": "Reminder: Upload your model in the next 48 hours",
      "comment_count": 70,
      "views": 0,
      "votes": -23
    },
    {
      "id": 25442,
      "title": "Welcome!",
      "comment_count": 11,
      "views": 1730,
      "votes": 31
    },
    {
      "id": 25432,
      "title": "Let's talk about the NDA",
      "comment_count": 8,
      "views": 3786,
      "votes": 15
    },
    {
      "id": 672166,
      "title": "Is the dataset available for academic use after the competition? (Late Submission Access)",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 27100,
      "title": "VGG/FCN solution (LB score ~0.987)",
      "comment_count": 28,
      "views": 2852,
      "votes": 74
    },
    {
      "id": 26202,
      "title": "Using Keras+TensorFlow to solve NCFM-Leadboard Top 5%",
      "comment_count": 51,
      "views": 6719,
      "votes": 51
    },
    {
      "id": 42521,
      "title": "Thanks to all the participants from TNC Australia",
      "comment_count": 0,
      "views": 0,
      "votes": 9
    },
    {
      "id": 25636,
      "title": "annotation tools",
      "comment_count": 10,
      "views": 756,
      "votes": 8
    },
    {
      "id": 30445,
      "title": "Deep Forest. 1.4 public log-loss with 100 samples",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 27089,
      "title": "Why can't the data be unzipped? ",
      "comment_count": 2,
      "views": 64,
      "votes": null
    },
    {
      "id": 37599,
      "title": "Boxes marked around fishes",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 30802,
      "title": "A SSD fish detector trained on imagenet",
      "comment_count": 10,
      "views": 0,
      "votes": 9
    },
    {
      "id": 31538,
      "title": "Solution Sharing and Congratulations",
      "comment_count": 66,
      "views": 0,
      "votes": 32
    },
    {
      "id": 31206,
      "title": "sample_submission_stg2.csv is wrong",
      "comment_count": 16,
      "views": 0,
      "votes": 7
    },
    {
      "id": 32300,
      "title": "Why isn't sample-submission in the leader board issue not resolved?",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 25439,
      "title": "Interesting solutions from the past.!",
      "comment_count": 4,
      "views": 1537,
      "votes": 52
    },
    {
      "id": 32031,
      "title": "Noob Problems",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 27866,
      "title": "Bounding box (object detection) using annotations",
      "comment_count": 3,
      "views": 257,
      "votes": 3
    },
    {
      "id": 31980,
      "title": "A Medium Article: A Brief History of CNNs in Image Segmentation",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 31607,
      "title": "Proposals for Paulo",
      "comment_count": 24,
      "views": 0,
      "votes": 6
    },
    {
      "id": 31232,
      "title": "Fish detection using a fully convolutional network",
      "comment_count": 5,
      "views": 1,
      "votes": 13
    },
    {
      "id": 31782,
      "title": "seems that Private Leaderboard is updated",
      "comment_count": 0,
      "views": 1,
      "votes": 1
    },
    {
      "id": 31211,
      "title": "Kaggle Admins: Please be more wiser and meticulous in the upcoming competitions.",
      "comment_count": 11,
      "views": 0,
      "votes": 55
    },
    {
      "id": 31542,
      "title": "Sample submission benchmark ranks so high in private board!",
      "comment_count": 12,
      "views": 0,
      "votes": 7
    },
    {
      "id": 31552,
      "title": "Should Sample Submission count as a submission for Ranking ?",
      "comment_count": 5,
      "views": 0,
      "votes": 6
    },
    {
      "id": 31586,
      "title": "Wrong problem statement",
      "comment_count": 6,
      "views": 0,
      "votes": 8
    },
    {
      "id": 31597,
      "title": "Why I Can't See Our Team Now",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 28215,
      "title": "Model training consistency in Keras",
      "comment_count": 6,
      "views": 202,
      "votes": null
    },
    {
      "id": 31465,
      "title": "You can't win if you don't submit",
      "comment_count": 8,
      "views": 0,
      "votes": 15
    },
    {
      "id": 31642,
      "title": "Why I can't  see my name on Leaderboard??",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 31543,
      "title": "How can Admin validate that participants are not modify their models since the LB shakeup so much?",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 31479,
      "title": "Will kaggle check model modification in stage2?",
      "comment_count": 9,
      "views": 0,
      "votes": 4
    },
    {
      "id": 31532,
      "title": "Is private LB gonna be immediately available?",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 31599,
      "title": "Kaggle please release the labeled test set",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 31563,
      "title": "What amazing result (1st place)",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31520,
      "title": "Submissions from non uploaded models",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 31554,
      "title": "Can someone explain the final decision to me?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 31292,
      "title": "The data in stg2 are so different from the training data",
      "comment_count": 12,
      "views": 0,
      "votes": 6
    },
    {
      "id": 31537,
      "title": "Regards the final ranking",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 31343,
      "title": "can we use 1% of the 2nd stage as public LB?",
      "comment_count": 0,
      "views": 0,
      "votes": 8
    },
    {
      "id": 31374,
      "title": "Kaggle admins: accidentally unselected one of my 2 submissions",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31325,
      "title": "To Admin:A question about the prediction.",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31338,
      "title": "What happened with the leader board",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 31321,
      "title": "rename stage2 file names of image_* to img_* with simple command line",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31287,
      "title": "why stage1 ranking changes after deadline of stage 1",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31204,
      "title": "a third of teams got wiped out in the LB",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 31279,
      "title": "patch script for correcting Stage2 submission image file name issue",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 31305,
      "title": "stg2 test data has only 12,153 images, instead of expected 13,154",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31207,
      "title": "question on stage2_test submission file format",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31233,
      "title": "Clarification needed on 2nd stage rule",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 31216,
      "title": "stage2 submission format",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31249,
      "title": "Why I can not find my models any more?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 30643,
      "title": "Time to discuss leaderboard shake up",
      "comment_count": 5,
      "views": 0,
      "votes": null
    },
    {
      "id": 31219,
      "title": "Will the score of stage2  be available only after the competition ends? ",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31247,
      "title": "Stage2 rules/clarification for those of us down to the wire.",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31217,
      "title": "4000 new images",
      "comment_count": 8,
      "views": 0,
      "votes": 1
    },
    {
      "id": 31215,
      "title": "Question + Confusion = Stage 2",
      "comment_count": 8,
      "views": 0,
      "votes": null
    },
    {
      "id": 31231,
      "title": "How to decompress test_stg2.7.z?",
      "comment_count": 3,
      "views": 2,
      "votes": null
    },
    {
      "id": 31240,
      "title": "Access Competition data in Kernel",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31222,
      "title": "Retrain model adding stg1 data",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 25565,
      "title": "annotated mouth and tail",
      "comment_count": 32,
      "views": 6555,
      "votes": 64
    },
    {
      "id": 31128,
      "title": "Suggestion: Torrent alternative for stage-2 download",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 31004,
      "title": "How does kaggle recognize multi-account user?",
      "comment_count": 4,
      "views": 0,
      "votes": 3
    },
    {
      "id": 31208,
      "title": "Unoffical Second Stage Clarification Thread",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 30873,
      "title": "Worse performance on cropped images?",
      "comment_count": 14,
      "views": 0,
      "votes": 5
    },
    {
      "id": 31123,
      "title": "I want to use the images from another source, as well from the data set can I do that?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31118,
      "title": "data scientist",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 31101,
      "title": "What exactly is the model?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 30971,
      "title": "where is the link to upload my model?",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 30872,
      "title": "What is multi-class logarithmic loss?",
      "comment_count": 2,
      "views": 0,
      "votes": -3
    },
    {
      "id": 30926,
      "title": "Looking for teammate",
      "comment_count": 7,
      "views": 0,
      "votes": null
    },
    {
      "id": 30540,
      "title": "A visualization of scores of our first submissions",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 30807,
      "title": "Can we use this dataset for research?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 30650,
      "title": "Any official details about model submission?",
      "comment_count": 2,
      "views": 0,
      "votes": 15
    },
    {
      "id": 30830,
      "title": "Is it allowed to use semi-supervised learning approach?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 30871,
      "title": "Training images for model",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 30144,
      "title": "Does anyone know of any Amazon AMI for deep learning on Windows?",
      "comment_count": 6,
      "views": 0,
      "votes": 0
    },
    {
      "id": 30813,
      "title": "Bounding box using keras",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 29575,
      "title": "Search team ( I worked almost 500 h on this project )",
      "comment_count": 8,
      "views": 0,
      "votes": null
    },
    {
      "id": 30355,
      "title": "complex deep network architecture end-to-end",
      "comment_count": 11,
      "views": 0,
      "votes": 5
    },
    {
      "id": 30714,
      "title": "Can we crop train images and use those cropped images for training the model?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 30054,
      "title": "How would I do transfer learning with Keras on the Resnet50 model?",
      "comment_count": 10,
      "views": 0,
      "votes": 1
    },
    {
      "id": 30456,
      "title": "final submission",
      "comment_count": 9,
      "views": 0,
      "votes": 1
    },
    {
      "id": 30280,
      "title": "This technique might be helpful since there are so many similar pictures.",
      "comment_count": 6,
      "views": 0,
      "votes": 6
    },
    {
      "id": 30563,
      "title": "Issue uploading submission files",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 30519,
      "title": "How to install EBImage package on Kaggle kernel",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 30338,
      "title": "How are we supposed to learn when there's no diversity",
      "comment_count": 2,
      "views": 0,
      "votes": 8
    },
    {
      "id": 30423,
      "title": "How to upload the model for first stage before second stage starts",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 30164,
      "title": "How to improve the performance of pretrained model?",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 30406,
      "title": "Preprocessing images to use vgg16 pretrained model",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 30317,
      "title": "How to deal with NoF?",
      "comment_count": 7,
      "views": 0,
      "votes": null
    },
    {
      "id": 29826,
      "title": "Model Submission",
      "comment_count": 8,
      "views": 0,
      "votes": 4
    },
    {
      "id": 30359,
      "title": "Has the timeline been updated for this competition?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 28917,
      "title": "How to upload model?",
      "comment_count": 3,
      "views": 0,
      "votes": 11
    },
    {
      "id": 29691,
      "title": "Any success with localization network?",
      "comment_count": 11,
      "views": 0,
      "votes": 1
    },
    {
      "id": 30234,
      "title": "Looking for teammates",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 30183,
      "title": "Amazon GPU AMI",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 28488,
      "title": "Things I've Tried...",
      "comment_count": 45,
      "views": 0,
      "votes": 50
    },
    {
      "id": 29988,
      "title": "#meta: how to filter searches by competition?",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 28621,
      "title": "Why is my CNN giving good accuracy on training and testing data (80-20 split of the total training data) but failing miserably in validation (unseen) data?",
      "comment_count": 8,
      "views": 0,
      "votes": 8
    },
    {
      "id": 29064,
      "title": "The most promising approach",
      "comment_count": 12,
      "views": 0,
      "votes": null
    },
    {
      "id": 29934,
      "title": "Validation Data Shuffling",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 29771,
      "title": "Issue with data augmentation",
      "comment_count": 7,
      "views": 0,
      "votes": 4
    },
    {
      "id": 29899,
      "title": "Question on cross validation",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 29773,
      "title": "Problems with GPU use on AWS Deep Learning AMI",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 29644,
      "title": "Evaluation problem please help ",
      "comment_count": 10,
      "views": 0,
      "votes": null
    },
    {
      "id": 29550,
      "title": "Fish Tag on Test Images",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 27600,
      "title": "private dataset should include ONLY new boats, and cryptohashes must be used",
      "comment_count": 8,
      "views": 669,
      "votes": 7
    },
    {
      "id": 29222,
      "title": "Path to test images",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 28407,
      "title": "How do you fight CNN overconfidence?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 29394,
      "title": "Nolearn / Lasagne and Log Loss Metrics",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 29258,
      "title": "Started need help  with convolutions !Tnks!",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 29284,
      "title": "loading images ",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 27219,
      "title": "can't figure out the format error in .csv file",
      "comment_count": 3,
      "views": 120,
      "votes": null
    },
    {
      "id": 29246,
      "title": "How to deal with the image which exists more that more fish speices?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 28823,
      "title": "Train set mixure",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 29131,
      "title": "Fish finder with example images",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 28972,
      "title": "matlab algorithm",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 29109,
      "title": "sample tensorflow code",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 27048,
      "title": "Single VGG16 pretrained for <1 logloss",
      "comment_count": 26,
      "views": 2644,
      "votes": 23
    },
    {
      "id": 28583,
      "title": "How to calculate correlation between submissions",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 29014,
      "title": "Python Script pixel by pixel to find fish",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 29016,
      "title": "Fish detector - python script",
      "comment_count": 0,
      "views": 0,
      "votes": -3
    },
    {
      "id": 28753,
      "title": "Searching for partner",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 28896,
      "title": "Can this problem be solved using Python?",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 28759,
      "title": "Issue  with  understanding  train_data.transpose((0, 3, 1, 2))",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 28442,
      "title": "Visualization tools to determine what features are being found in CNNs?",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 28691,
      "title": "Need suggestion in selecting images for training",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 28592,
      "title": "Issue  understanding  the   3 d format  and maipulate   them",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 28306,
      "title": "Simple validation strategy",
      "comment_count": 4,
      "views": 84,
      "votes": 4
    },
    {
      "id": 27697,
      "title": "bbox strategies",
      "comment_count": 6,
      "views": 552,
      "votes": 4
    },
    {
      "id": 28150,
      "title": "Unified effort to relabel the Training set.",
      "comment_count": 17,
      "views": 536,
      "votes": 25
    },
    {
      "id": 25988,
      "title": "Dealing with data imbalance",
      "comment_count": 3,
      "views": 402,
      "votes": 0
    },
    {
      "id": 28390,
      "title": "How to Tune Hyper-parameter in Keras",
      "comment_count": 7,
      "views": 0,
      "votes": null
    },
    {
      "id": 28439,
      "title": "The dream train and loss curves .... or so I thought",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 27992,
      "title": "what are your computing specs + training speed?",
      "comment_count": 4,
      "views": 228,
      "votes": null
    },
    {
      "id": 28225,
      "title": "Unified effort to track sequences.",
      "comment_count": 4,
      "views": 189,
      "votes": 5
    },
    {
      "id": 27677,
      "title": "How to determine whether to submit model",
      "comment_count": 12,
      "views": 562,
      "votes": null
    },
    {
      "id": 28078,
      "title": "Team",
      "comment_count": 10,
      "views": 369,
      "votes": 2
    },
    {
      "id": 26492,
      "title": "Looking for a team mate",
      "comment_count": 4,
      "views": 493,
      "votes": null
    },
    {
      "id": 27911,
      "title": "How to interpret the loss results?",
      "comment_count": 5,
      "views": 348,
      "votes": null
    },
    {
      "id": 27858,
      "title": "An article about this Overexplotation of fish stocks",
      "comment_count": 0,
      "views": 81,
      "votes": 2
    },
    {
      "id": 27810,
      "title": "Data for Thesis Project?",
      "comment_count": 2,
      "views": 207,
      "votes": 1
    },
    {
      "id": 27361,
      "title": "Models completely fail on new boats",
      "comment_count": 18,
      "views": 1270,
      "votes": 16
    },
    {
      "id": 27757,
      "title": "Looking for a teamate good at fish",
      "comment_count": 7,
      "views": 424,
      "votes": null
    },
    {
      "id": 27745,
      "title": "Free AWS g2 or p2 instances until end of month!",
      "comment_count": 8,
      "views": 530,
      "votes": 1
    },
    {
      "id": 27719,
      "title": "NoF Class",
      "comment_count": 3,
      "views": 332,
      "votes": 1
    },
    {
      "id": 25460,
      "title": "Step 1: Locate the fish in the image",
      "comment_count": 4,
      "views": 1317,
      "votes": 18
    },
    {
      "id": 27825,
      "title": "I have cheated but I have learned my lesson",
      "comment_count": 0,
      "views": 447,
      "votes": -11
    },
    {
      "id": 27399,
      "title": "suggestions of validation strategy?",
      "comment_count": 1,
      "views": 276,
      "votes": 2
    },
    {
      "id": 26415,
      "title": "fish image 00568?",
      "comment_count": 21,
      "views": 1866,
      "votes": 2
    },
    {
      "id": 27059,
      "title": "Improved Cross Validation Schemes, Anyone?",
      "comment_count": 3,
      "views": 440,
      "votes": 4
    },
    {
      "id": 27565,
      "title": "Fish recognition is much more challenging than face recognition",
      "comment_count": 0,
      "views": 277,
      "votes": 0
    },
    {
      "id": 27442,
      "title": "Which amazon instance to take for Neural network?",
      "comment_count": 7,
      "views": 464,
      "votes": 1
    },
    {
      "id": 27427,
      "title": "IOU for bounding box fish detection",
      "comment_count": 3,
      "views": 518,
      "votes": null
    },
    {
      "id": 27471,
      "title": "No fish detection == overfitting ?",
      "comment_count": 2,
      "views": 293,
      "votes": null
    },
    {
      "id": 27456,
      "title": "can caffe be used for solving this problem?",
      "comment_count": 1,
      "views": 149,
      "votes": null
    },
    {
      "id": 27438,
      "title": "Dataset too small",
      "comment_count": 2,
      "views": 321,
      "votes": null
    },
    {
      "id": 27393,
      "title": "sha512sum instead of model uploads?",
      "comment_count": 0,
      "views": 87,
      "votes": 1
    },
    {
      "id": 26676,
      "title": "Yet another topic about ambiguous data - ALB",
      "comment_count": 6,
      "views": 813,
      "votes": 5
    },
    {
      "id": 27346,
      "title": "pca and tsne",
      "comment_count": 7,
      "views": 382,
      "votes": 1
    },
    {
      "id": 27370,
      "title": "Finding LAG",
      "comment_count": 0,
      "views": 171,
      "votes": 1
    },
    {
      "id": 27338,
      "title": "Best single model score without cropping",
      "comment_count": 0,
      "views": 227,
      "votes": 0
    },
    {
      "id": 27212,
      "title": "Creative Data Augmentation using Generators for Keras",
      "comment_count": 0,
      "views": 225,
      "votes": 7
    },
    {
      "id": 27202,
      "title": "openCV filter cascade object detection ",
      "comment_count": 6,
      "views": 365,
      "votes": null
    },
    {
      "id": 27195,
      "title": "Are there some training samples with wrong labels?",
      "comment_count": 0,
      "views": 125,
      "votes": 1
    },
    {
      "id": 26979,
      "title": "How to best use the bounding boxes",
      "comment_count": 1,
      "views": 384,
      "votes": null
    },
    {
      "id": 27138,
      "title": "Guide to kaggle docker",
      "comment_count": 1,
      "views": 221,
      "votes": null
    },
    {
      "id": 26022,
      "title": "More then one fish again",
      "comment_count": 4,
      "views": 1041,
      "votes": 9
    },
    {
      "id": 27003,
      "title": "Fish detection using deep learning heatmaps.",
      "comment_count": 3,
      "views": 678,
      "votes": 6
    },
    {
      "id": 27021,
      "title": "In sample testing results do not match out sample testing result.",
      "comment_count": 5,
      "views": 235,
      "votes": null
    },
    {
      "id": 26944,
      "title": "any one using sk image python ",
      "comment_count": 1,
      "views": 227,
      "votes": null
    },
    {
      "id": 26962,
      "title": "Importing data",
      "comment_count": 0,
      "views": 136,
      "votes": 0
    },
    {
      "id": 26910,
      "title": "any post-processing?",
      "comment_count": 2,
      "views": 280,
      "votes": null
    },
    {
      "id": 26375,
      "title": "Upload Model",
      "comment_count": 2,
      "views": 285,
      "votes": null
    },
    {
      "id": 26815,
      "title": "Visualizing/diagnosing neural network.",
      "comment_count": 2,
      "views": 470,
      "votes": 8
    },
    {
      "id": 25737,
      "title": "What Computer Equipment Required for this Competition?",
      "comment_count": 19,
      "views": 1897,
      "votes": 0
    },
    {
      "id": 26799,
      "title": "How to save time for training process?",
      "comment_count": 1,
      "views": 223,
      "votes": 1
    },
    {
      "id": 25473,
      "title": "Questions about 2nd stage of competition",
      "comment_count": 39,
      "views": 3993,
      "votes": 21
    },
    {
      "id": 26770,
      "title": "Extreme Fish",
      "comment_count": 0,
      "views": 176,
      "votes": 1
    },
    {
      "id": 26636,
      "title": "Train-Test split",
      "comment_count": 0,
      "views": 226,
      "votes": 0
    },
    {
      "id": 25457,
      "title": "That's not albacore !?",
      "comment_count": 28,
      "views": 3213,
      "votes": 2
    },
    {
      "id": 26480,
      "title": "How to delete kernel?",
      "comment_count": 0,
      "views": 114,
      "votes": 4
    },
    {
      "id": 26481,
      "title": "Program Language Limitation",
      "comment_count": 2,
      "views": 309,
      "votes": null
    },
    {
      "id": 26425,
      "title": "How the competitions work while codes are there?",
      "comment_count": 7,
      "views": 597,
      "votes": null
    },
    {
      "id": 26371,
      "title": "Keras thinks I have 9 classes",
      "comment_count": 1,
      "views": 247,
      "votes": null
    },
    {
      "id": 26204,
      "title": "Permission about data set usage",
      "comment_count": 0,
      "views": 150,
      "votes": 0
    },
    {
      "id": 26048,
      "title": "To Align or Not To Align !!!",
      "comment_count": 7,
      "views": 834,
      "votes": 3
    },
    {
      "id": 25749,
      "title": "How many kagglers fish IRL?",
      "comment_count": 5,
      "views": 658,
      "votes": 8
    },
    {
      "id": 26110,
      "title": "negated average binary loss to probability per class",
      "comment_count": 0,
      "views": 131,
      "votes": 0
    },
    {
      "id": 26088,
      "title": "Request for Admins: Can we get info on the camera field of view/focal length? ",
      "comment_count": 0,
      "views": 201,
      "votes": 0
    },
    {
      "id": 25909,
      "title": "getting started with image classification",
      "comment_count": 2,
      "views": 563,
      "votes": 0
    },
    {
      "id": 25818,
      "title": "Clustering+Mask+Classify Workflow",
      "comment_count": 8,
      "views": 687,
      "votes": 3
    },
    {
      "id": 25833,
      "title": "Train and Test Set Distributions",
      "comment_count": 0,
      "views": 292,
      "votes": 2
    },
    {
      "id": 25548,
      "title": "Softmax Log Loss and Caffe",
      "comment_count": 12,
      "views": 1167,
      "votes": 0
    },
    {
      "id": 25733,
      "title": "Cross Validation",
      "comment_count": 1,
      "views": 333,
      "votes": null
    },
    {
      "id": 25729,
      "title": "Duplicates in Test Data",
      "comment_count": 5,
      "views": 309,
      "votes": 1
    },
    {
      "id": 25659,
      "title": "Press about this competition",
      "comment_count": 0,
      "views": 386,
      "votes": 6
    },
    {
      "id": 25627,
      "title": "1.64611",
      "comment_count": 6,
      "views": 521,
      "votes": 1
    },
    {
      "id": 25476,
      "title": "Human Performance",
      "comment_count": 6,
      "views": 1063,
      "votes": 8
    },
    {
      "id": 25600,
      "title": "The images are JPEGs",
      "comment_count": 0,
      "views": 290,
      "votes": 0
    },
    {
      "id": 25493,
      "title": "Lat's discuss possible methods here!",
      "comment_count": 1,
      "views": 803,
      "votes": 14
    },
    {
      "id": 25478,
      "title": "Using RStudio Server for Uploading Kaggle DataSet",
      "comment_count": 22,
      "views": 794,
      "votes": 0
    },
    {
      "id": 25537,
      "title": "Proposed improvement to log loss function",
      "comment_count": 6,
      "views": 496,
      "votes": null
    },
    {
      "id": 25463,
      "title": "More than one fish specie shown on the image.",
      "comment_count": 2,
      "views": 799,
      "votes": 2
    },
    {
      "id": 25444,
      "title": "\"This leaderboard is calculated on all of the test data.\"",
      "comment_count": 2,
      "views": 869,
      "votes": 8
    },
    {
      "id": 25431,
      "title": "Is this the First Kaggle Competition with a Warning?",
      "comment_count": 1,
      "views": 945,
      "votes": 5
    },
    {
      "id": 25453,
      "title": "vgg_16_initial_model",
      "comment_count": 2,
      "views": 534,
      "votes": 2
    },
    {
      "id": 25487,
      "title": "What's up with the score ? ",
      "comment_count": 4,
      "views": 528,
      "votes": null
    },
    {
      "id": 25469,
      "title": "Guidance on image classification",
      "comment_count": 2,
      "views": 887,
      "votes": null
    },
    {
      "id": 25465,
      "title": "Absolute Novice - Getting started ",
      "comment_count": 0,
      "views": 332,
      "votes": 0
    },
    {
      "id": 25427,
      "title": "pretrained networks",
      "comment_count": 1,
      "views": 815,
      "votes": 1
    },
    {
      "id": 25435,
      "title": "Team up",
      "comment_count": 2,
      "views": 431,
      "votes": 1
    }
  ],
  "errors": []
}