{
  "total": 183,
  "topics": [
    {
      "id": 146593,
      "title": "New to Machine Learning or Kaggle? Come say Hi!",
      "comment_count": 22,
      "views": 0,
      "votes": 8
    },
    {
      "id": 170707,
      "title": "Leaderboard is Finalized - Congratulations to our Winners!",
      "comment_count": 3,
      "views": 0,
      "votes": 16
    },
    {
      "id": 147182,
      "title": "Important changes in the evaluation (and apologies)",
      "comment_count": 11,
      "views": 0,
      "votes": 34
    },
    {
      "id": 155392,
      "title": "A few tips for easy start",
      "comment_count": 48,
      "views": 0,
      "votes": 157
    },
    {
      "id": 252983,
      "title": "Recent topics",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 168548,
      "title": "1st place solution",
      "comment_count": 82,
      "views": 0,
      "votes": 175
    },
    {
      "id": 159515,
      "title": "List of Important Research Papers for Competition",
      "comment_count": 13,
      "views": 0,
      "votes": 39
    },
    {
      "id": 155167,
      "title": "Yet another starter [training + inference]",
      "comment_count": 18,
      "views": 0,
      "votes": 23
    },
    {
      "id": 168702,
      "title": "[34th place] My first silver solution :P",
      "comment_count": 16,
      "views": 0,
      "votes": 25
    },
    {
      "id": 173489,
      "title": "[ABBA McCandless] - 2nd place solution code",
      "comment_count": 3,
      "views": 0,
      "votes": 13
    },
    {
      "id": 168611,
      "title": " congrats to Guanshuo Xu! and [14th place] solution ",
      "comment_count": 13,
      "views": 0,
      "votes": 47
    },
    {
      "id": 168870,
      "title": "3rd place solution",
      "comment_count": 3,
      "views": 0,
      "votes": 45
    },
    {
      "id": 168542,
      "title": "Single Effnet-B0 Private LB 0.921: How to Modify Effnet Architecture",
      "comment_count": 17,
      "views": 0,
      "votes": 74
    },
    {
      "id": 168673,
      "title": "AMA with the new king of competitions: Guanshuo Xu | CTDS.Show",
      "comment_count": 3,
      "views": 0,
      "votes": 11
    },
    {
      "id": 169448,
      "title": "0.927 Private LB single model (1-fold) B0-inspired architecture (late sub)",
      "comment_count": 6,
      "views": 0,
      "votes": 15
    },
    {
      "id": 170537,
      "title": "LB cleaned up, but still remain Deleted user?",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 168510,
      "title": "Private LB 13th Solution - so near yet so far :)",
      "comment_count": 7,
      "views": 0,
      "votes": 41
    },
    {
      "id": 168798,
      "title": "Finalizing Leaderboard Soon!",
      "comment_count": 2,
      "views": 0,
      "votes": 11
    },
    {
      "id": 168608,
      "title": "[9th place] Short summary",
      "comment_count": 20,
      "views": 0,
      "votes": 57
    },
    {
      "id": 168771,
      "title": "18th Place Brief Solution Overview",
      "comment_count": 10,
      "views": 0,
      "votes": 50
    },
    {
      "id": 170943,
      "title": "Did Anyone Have Success Using Accumulated Gradients?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 168537,
      "title": "[4th place] Definitely it's overfitting to the public LB. (0.948->0.930)",
      "comment_count": 24,
      "views": 0,
      "votes": 57
    },
    {
      "id": 168519,
      "title": "[8th Place] Brief solution.",
      "comment_count": 21,
      "views": 0,
      "votes": 37
    },
    {
      "id": 168546,
      "title": "[ABBA McCandless] - 2nd place solution overview ",
      "comment_count": 17,
      "views": 0,
      "votes": 75
    },
    {
      "id": 168507,
      "title": "[12th place] Solution Overview",
      "comment_count": 18,
      "views": 0,
      "votes": 51
    },
    {
      "id": 168509,
      "title": "[31st place] with public kernel",
      "comment_count": 10,
      "views": 0,
      "votes": 23
    },
    {
      "id": 168410,
      "title": "Submission < 10 but in Top 50 - Interesting Solution Coming!",
      "comment_count": 30,
      "views": 0,
      "votes": 22
    },
    {
      "id": 168580,
      "title": "Solutions Collection (To be Updated)",
      "comment_count": 3,
      "views": 0,
      "votes": 10
    },
    {
      "id": 168555,
      "title": "Congratulations to you all ",
      "comment_count": 4,
      "views": 0,
      "votes": 19
    },
    {
      "id": 168496,
      "title": "Done my final submissions and good luck for everyone!",
      "comment_count": 3,
      "views": 0,
      "votes": 13
    },
    {
      "id": 168528,
      "title": "Wow it looks like we will have a new kaggle number one!",
      "comment_count": 3,
      "views": 0,
      "votes": 15
    },
    {
      "id": 169447,
      "title": "A single model inference (public/private 0.940/0.929)",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 168525,
      "title": "survey of TPU usage",
      "comment_count": 13,
      "views": 0,
      "votes": 4
    },
    {
      "id": 168572,
      "title": " [74th place] First Medal - How I missed the Silver Medal",
      "comment_count": 2,
      "views": 0,
      "votes": 9
    },
    {
      "id": 167652,
      "title": "Model Error Analysis",
      "comment_count": 11,
      "views": 0,
      "votes": 12
    },
    {
      "id": 168506,
      "title": "It's over! What did we learn?",
      "comment_count": 1,
      "views": 0,
      "votes": 8
    },
    {
      "id": 168107,
      "title": "Will there be a shakeup?",
      "comment_count": 16,
      "views": 0,
      "votes": 2
    },
    {
      "id": 168515,
      "title": "why using ycbcr is not recommended?",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 168091,
      "title": "All the best!",
      "comment_count": 0,
      "views": 0,
      "votes": 6
    },
    {
      "id": 168440,
      "title": "Done!",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 167033,
      "title": "Screw DCT",
      "comment_count": 3,
      "views": 0,
      "votes": 11
    },
    {
      "id": 168316,
      "title": "Can someone help me with this error ?",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 168534,
      "title": "Congratulations!",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 167300,
      "title": "Code in TF for YCbCr conversion",
      "comment_count": 3,
      "views": 0,
      "votes": 4
    },
    {
      "id": 164645,
      "title": "Crossing the 0.921 baseline!",
      "comment_count": 42,
      "views": 0,
      "votes": 16
    },
    {
      "id": 165411,
      "title": "Feedbacks, submissions and associated conference",
      "comment_count": 10,
      "views": 0,
      "votes": 17
    },
    {
      "id": 166331,
      "title": "Single vs Ensemble Scores.",
      "comment_count": 11,
      "views": 0,
      "votes": 2
    },
    {
      "id": 164319,
      "title": "Ideas for augmentation, network design, etc",
      "comment_count": 20,
      "views": 0,
      "votes": 22
    },
    {
      "id": 165338,
      "title": "ValueError: Input contains NaN, infinity or a value too large for dtype('float32')",
      "comment_count": 13,
      "views": 0,
      "votes": 8
    },
    {
      "id": 165994,
      "title": "TPU High LR doesn't seem to work very well?",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 164256,
      "title": "Metric questions",
      "comment_count": 7,
      "views": 0,
      "votes": 15
    },
    {
      "id": 162154,
      "title": "It's unfair !",
      "comment_count": 20,
      "views": 0,
      "votes": 10
    },
    {
      "id": 165651,
      "title": "Gradient accumulation - How to speed up convergence",
      "comment_count": 3,
      "views": 0,
      "votes": 13
    },
    {
      "id": 165395,
      "title": "A Return to DCT",
      "comment_count": 6,
      "views": 0,
      "votes": 13
    },
    {
      "id": 166685,
      "title": "[Done!]Looking for a Team!!",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 166253,
      "title": "EfficientNet- GPU or Tensors",
      "comment_count": 5,
      "views": 0,
      "votes": null
    },
    {
      "id": 165544,
      "title": "Keras Labelsmoothing Implementation",
      "comment_count": 4,
      "views": 0,
      "votes": 7
    },
    {
      "id": 166185,
      "title": "How many epochs are good enough or necessary to get stable score?",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 165070,
      "title": "Low Single Fold CV/LB with Keras",
      "comment_count": 4,
      "views": 0,
      "votes": 3
    },
    {
      "id": 165256,
      "title": "Advantage of noisy-student",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 165081,
      "title": "[SOLVED] How to use Colab TPUs for training?",
      "comment_count": 8,
      "views": 0,
      "votes": 3
    },
    {
      "id": 165488,
      "title": "ZeroDivisionError: float division by zero",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 165776,
      "title": "What should be tf.image.rgb_to_yuv input range?",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 165125,
      "title": "ALASKA-3 can be a two-player game challenge ....",
      "comment_count": 4,
      "views": 0,
      "votes": 14
    },
    {
      "id": 165590,
      "title": "does Pseudo-Label work in this game？",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 164184,
      "title": "understanding from scratch with code: UERD (uniform embedding revisited distortion)",
      "comment_count": 1,
      "views": 0,
      "votes": 11
    },
    {
      "id": 164828,
      "title": "Learning Rich Features for Image Manipulation Detection",
      "comment_count": 1,
      "views": 0,
      "votes": 10
    },
    {
      "id": 163623,
      "title": "hello Colab TPU User's",
      "comment_count": 8,
      "views": 0,
      "votes": 3
    },
    {
      "id": 165130,
      "title": "Public Kaggle Dataset on Personal TPU?",
      "comment_count": 7,
      "views": 0,
      "votes": 3
    },
    {
      "id": 164996,
      "title": "Benefits of TFRecord Dataset?  ",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 162177,
      "title": "Why does Efficientnet work?",
      "comment_count": 14,
      "views": 0,
      "votes": 9
    },
    {
      "id": 162936,
      "title": "An estimation of public/private LB",
      "comment_count": 3,
      "views": 0,
      "votes": 11
    },
    {
      "id": 162507,
      "title": "Request to have more digits on LB.",
      "comment_count": 10,
      "views": 0,
      "votes": 0
    },
    {
      "id": 164766,
      "title": "Any alternative to EfficientNets? ",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 164633,
      "title": "Problem when using TPU, MXU. ",
      "comment_count": 7,
      "views": 0,
      "votes": null
    },
    {
      "id": 164641,
      "title": "Kernel Performance Issue",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 164155,
      "title": "understanding from scratch with code: J-MiPOD (Minimizing the Power of Optimal Detector)",
      "comment_count": 1,
      "views": 0,
      "votes": 11
    },
    {
      "id": 163535,
      "title": "How to Implement Custom Metric on TPU?",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 163780,
      "title": "What happens next?",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 164604,
      "title": "TPU Time limit.",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 163629,
      "title": "Using Kaggle GPU for training",
      "comment_count": 3,
      "views": 0,
      "votes": 4
    },
    {
      "id": 163073,
      "title": "BitMix: Data Augmentation for Image Steganalysis",
      "comment_count": 6,
      "views": 0,
      "votes": 21
    },
    {
      "id": 164351,
      "title": "Did anyone get any significant progress with just metadata?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 163992,
      "title": "understanding from scratch with code: J-UNIWARD ( Universal Wavelet Relative Distortion)",
      "comment_count": 6,
      "views": 0,
      "votes": 14
    },
    {
      "id": 163902,
      "title": "one idea to train efficient-b7 on large 512x512 image",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 163683,
      "title": "competition metric",
      "comment_count": 7,
      "views": 0,
      "votes": 7
    },
    {
      "id": 163763,
      "title": "Competition Metrics with Tensorflow/Keras",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 163962,
      "title": "tfrecords of 30K images",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 162957,
      "title": "TPU Memory Problem",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 162982,
      "title": "I am beginner can you tell me where to begin?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 163841,
      "title": "which is better training on TPU or GPU , Images or TfRecords ? What does the GrandMasters Suggest On this ?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 161439,
      "title": "Overfitting ?",
      "comment_count": 16,
      "views": 0,
      "votes": 8
    },
    {
      "id": 163675,
      "title": "Is it important to keep a cover-stego pair in a minibatch?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 162338,
      "title": "Can we open 4th digit on the lb?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 150359,
      "title": "RGB, YUV or DCT which one should be used?",
      "comment_count": 34,
      "views": 0,
      "votes": 31
    },
    {
      "id": 162341,
      "title": "Colab TPU Error: UnavailableError: Socket closed",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 163553,
      "title": "Can You get a good score without using all images ? ",
      "comment_count": 1,
      "views": 0,
      "votes": 0
    },
    {
      "id": 163547,
      "title": "PyTorch Image Models",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 161861,
      "title": "Image quality 95 most difficult to detect?",
      "comment_count": 4,
      "views": 0,
      "votes": 3
    },
    {
      "id": 163546,
      "title": "Facebook codebase PYCLS",
      "comment_count": 0,
      "views": 0,
      "votes": -2
    },
    {
      "id": 162470,
      "title": "Single Channel Image",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 160375,
      "title": "How can I save a model?",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 161665,
      "title": "denoising approach",
      "comment_count": 13,
      "views": 0,
      "votes": 11
    },
    {
      "id": 160315,
      "title": "New to kaggle",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 161878,
      "title": "What is your setup?",
      "comment_count": 2,
      "views": 0,
      "votes": 7
    },
    {
      "id": 155814,
      "title": "JUNIWARD is the most difficult one to detect ?",
      "comment_count": 19,
      "views": 0,
      "votes": 19
    },
    {
      "id": 161986,
      "title": "Denoising open-source resources",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 148162,
      "title": "Best single model",
      "comment_count": 97,
      "views": 0,
      "votes": 39
    },
    {
      "id": 156471,
      "title": "How to Iterate/Experiment fast(er)?",
      "comment_count": 8,
      "views": 0,
      "votes": 1
    },
    {
      "id": 161905,
      "title": "Suggestions to the Kaggle team. ",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 161312,
      "title": "Batch norm problems",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 161388,
      "title": "Mixed precision training using PyTorch 1.6",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 161585,
      "title": "[HELP NEEDED] TF Custom Training Loop on TPU",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 161465,
      "title": "input files are just disappeared from /kaggle/input",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 161377,
      "title": "how to submit the kernel?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 161251,
      "title": "os.environ",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 159612,
      "title": "TPU running time is 5 times more than GPU",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 158308,
      "title": "Corrupt images?",
      "comment_count": 7,
      "views": 0,
      "votes": 9
    },
    {
      "id": 159558,
      "title": "Did anyone try Adversarial Training ?",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 157435,
      "title": "Is it possible to use mixup in this competition? ",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 159996,
      "title": "40+ pretrained models on pytorch",
      "comment_count": 9,
      "views": 0,
      "votes": 3
    },
    {
      "id": 156481,
      "title": "How to Download Large Datasets for Beginners?",
      "comment_count": 9,
      "views": 0,
      "votes": 2
    },
    {
      "id": 158260,
      "title": "Anyone tried to learn Cover vs Juniward model? ",
      "comment_count": 7,
      "views": 0,
      "votes": 3
    },
    {
      "id": 159159,
      "title": "can you please help to get the metric in tensorflow",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 151428,
      "title": "Anyone Tried Google  Colab?",
      "comment_count": 16,
      "views": 0,
      "votes": 7
    },
    {
      "id": 146595,
      "title": "Looking for a Team Megathread",
      "comment_count": 29,
      "views": 0,
      "votes": 6
    },
    {
      "id": 155422,
      "title": "Has anyone tried mixed-precision training?",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 146597,
      "title": "External Discussion Thread",
      "comment_count": 21,
      "views": 0,
      "votes": 1
    },
    {
      "id": 157659,
      "title": "Trouble uploading data to colab. Help?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 147039,
      "title": "Alaska Winning Solution",
      "comment_count": 5,
      "views": 0,
      "votes": 27
    },
    {
      "id": 148986,
      "title": "How do I download dataset to Colab?",
      "comment_count": 4,
      "views": 0,
      "votes": 6
    },
    {
      "id": 152524,
      "title": "Help ! UnavailableError: Socket closed",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 158006,
      "title": "Loss Decreases to `nan` after ~14 epochs in fp16",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 155536,
      "title": "Anyone know why my images aren't showing?",
      "comment_count": 3,
      "views": 0,
      "votes": null
    },
    {
      "id": 156163,
      "title": "Some findings from analysing DCT table changes during embedding",
      "comment_count": 1,
      "views": 0,
      "votes": 12
    },
    {
      "id": 155821,
      "title": "Justifying multiclass.",
      "comment_count": 6,
      "views": 0,
      "votes": 22
    },
    {
      "id": 156523,
      "title": "Summary on Steganalysis",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 153247,
      "title": "Pre-training for Computer Vision ",
      "comment_count": 1,
      "views": 0,
      "votes": 8
    },
    {
      "id": 153166,
      "title": "Possible Multi-Stage Pipeline?",
      "comment_count": 3,
      "views": 0,
      "votes": 4
    },
    {
      "id": 154854,
      "title": "Ask to TPU users",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 154898,
      "title": "Combining Ycbcr Conversion with tf.data.Dataset",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 151631,
      "title": "Error using KaggleDatasets().get_gcs_path()",
      "comment_count": 15,
      "views": 0,
      "votes": 7
    },
    {
      "id": 154843,
      "title": "Take into account the cost functions of the embedding schemes?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 153010,
      "title": "Splitting data based on Quality factor and Stego Scheme",
      "comment_count": 3,
      "views": 0,
      "votes": 8
    },
    {
      "id": 153448,
      "title": "Pytorch Hacks for Speedup",
      "comment_count": 1,
      "views": 0,
      "votes": 19
    },
    {
      "id": 148919,
      "title": "Anyone tried this as multiclass classification problem ?",
      "comment_count": 3,
      "views": 0,
      "votes": 13
    },
    {
      "id": 146622,
      "title": "Alaska2: Intro + Literature Review + External Data + Networks + Relevant Githubs",
      "comment_count": 17,
      "views": 0,
      "votes": 60
    },
    {
      "id": 153616,
      "title": "Prediction Output Format",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 146950,
      "title": "JMiPOD, JUNIWARD, UERD",
      "comment_count": 8,
      "views": 0,
      "votes": 30
    },
    {
      "id": 148090,
      "title": "New Topic : Image Steganalysis",
      "comment_count": 3,
      "views": 0,
      "votes": 23
    },
    {
      "id": 151769,
      "title": "ways to learn the NN that does not fit in GPU memory with reasonable minibatch.",
      "comment_count": 6,
      "views": 0,
      "votes": null
    },
    {
      "id": 151526,
      "title": "Network's loss not decreasing",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 150592,
      "title": "Checking My Understanding ",
      "comment_count": 8,
      "views": 0,
      "votes": 0
    },
    {
      "id": 150561,
      "title": "Anyone looking at it from a One-Shot Learning Perspective?",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 149689,
      "title": "In TPU, can't I use 'class_weight' in 'model.fit()' ?",
      "comment_count": 4,
      "views": 0,
      "votes": 6
    },
    {
      "id": 151445,
      "title": "Is there a way to determine the jpeg image compression rate?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 150943,
      "title": "more image processing",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 150596,
      "title": "JPEG Quality Factor",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 150268,
      "title": "Index Error in weighted AUC?",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 150622,
      "title": "Distinguish features ",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 150520,
      "title": "Memory errors ! How are you fitting the whole data ?",
      "comment_count": 1,
      "views": 0,
      "votes": 0
    },
    {
      "id": 150300,
      "title": "Is there any use of seeing weighted AUC on a batch?",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 149238,
      "title": "How does Data Augmentation and transformations affects on Steganalysis?",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 146607,
      "title": "Sounds like a really interesting competition",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 147494,
      "title": "Impact of steganography: do not look only at pixel difference",
      "comment_count": 3,
      "views": 0,
      "votes": 43
    },
    {
      "id": 147057,
      "title": "Methods To Approach The Problem And Their Inferences",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 146699,
      "title": "If you want more training data ",
      "comment_count": 6,
      "views": 0,
      "votes": 8
    },
    {
      "id": 149179,
      "title": "Plethora of pytorch pretrained models",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 148757,
      "title": "Image Classification Solutions",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 148752,
      "title": "Share great guide how to use TPU",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 148653,
      "title": "TPU Cannot use tf.py_function?? Or am I not writing the right code?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 147947,
      "title": "Problem getting data through Kaggle API",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 146610,
      "title": "Paper worth mentioning",
      "comment_count": 13,
      "views": 0,
      "votes": 43
    },
    {
      "id": 146989,
      "title": "Discrepancy between accuracy and the competition metric",
      "comment_count": 9,
      "views": 0,
      "votes": 12
    },
    {
      "id": 147751,
      "title": "Pytorch Transfer Learning Baseline",
      "comment_count": 6,
      "views": 0,
      "votes": 4
    },
    {
      "id": 147367,
      "title": "[Info]: Keras (TF 2) Practice intro from François Chollet",
      "comment_count": 0,
      "views": 0,
      "votes": 7
    },
    {
      "id": 146604,
      "title": "Welcome !",
      "comment_count": 3,
      "views": 0,
      "votes": 15
    },
    {
      "id": 148022,
      "title": "Unique Solution and comprehensive understanding of Data",
      "comment_count": 1,
      "views": 0,
      "votes": -3
    },
    {
      "id": 147153,
      "title": "Understanding the LB",
      "comment_count": 1,
      "views": 0,
      "votes": 6
    },
    {
      "id": 147050,
      "title": "Suggestions on solving Memory allocation Problem",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 147120,
      "title": "Changing the metrics of the competition while it's early",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 146668,
      "title": "Deep Learning or Traditional Steganalysis?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 146669,
      "title": "Deep Learning or Traditional Steganalysis?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    }
  ],
  "errors": []
}