{
  "total": 132,
  "topics": [
    {
      "id": 186493,
      "title": "Solutions of top-ranked teams",
      "comment_count": 6,
      "views": 0,
      "votes": 11
    },
    {
      "id": 183732,
      "title": "ILR workshop at ECCV'20",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 175627,
      "title": "Congrats to the Winners! Don't forget to submit your documentation!",
      "comment_count": 7,
      "views": 0,
      "votes": 9
    },
    {
      "id": 163341,
      "title": "Your First Submission (Baseline Example)",
      "comment_count": 14,
      "views": 0,
      "votes": 43
    },
    {
      "id": 163038,
      "title": "External Data Thread",
      "comment_count": 26,
      "views": 0,
      "votes": 0
    },
    {
      "id": 165400,
      "title": "Scoring Script",
      "comment_count": 6,
      "views": 0,
      "votes": 13
    },
    {
      "id": 163037,
      "title": "Looking for a Team Megathread",
      "comment_count": 80,
      "views": 0,
      "votes": 7
    },
    {
      "id": 163350,
      "title": "Your Second Submission (model training/exporting example)",
      "comment_count": 18,
      "views": 0,
      "votes": 42
    },
    {
      "id": 163027,
      "title": "Welcome to Landmark Retrieval 2020",
      "comment_count": 14,
      "views": 0,
      "votes": 22
    },
    {
      "id": 163036,
      "title": "New to Machine Learning or to Kaggle? ",
      "comment_count": 6,
      "views": 0,
      "votes": 2
    },
    {
      "id": 175472,
      "title": "9th place solution overview",
      "comment_count": 17,
      "views": 0,
      "votes": 39
    },
    {
      "id": 176037,
      "title": "1st place solution summary",
      "comment_count": 110,
      "views": 0,
      "votes": 142
    },
    {
      "id": 163628,
      "title": "Any difference between this dataset and GLDv2 ?",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 176151,
      "title": "5th place solution write-up ",
      "comment_count": 14,
      "views": 0,
      "votes": 37
    },
    {
      "id": 175306,
      "title": "I beat baseline.277 with EFN on Keras. Limited time and resource stop me further.",
      "comment_count": 8,
      "views": 0,
      "votes": 14
    },
    {
      "id": 175343,
      "title": "Finetuning the baseline - silver medal solution [0.285 public LB]",
      "comment_count": 5,
      "views": 0,
      "votes": 14
    },
    {
      "id": 177078,
      "title": "2nd place solution summary",
      "comment_count": 16,
      "views": 0,
      "votes": 35
    },
    {
      "id": 175378,
      "title": "4th place solution and experience sharing",
      "comment_count": 26,
      "views": 0,
      "votes": 58
    },
    {
      "id": 177162,
      "title": "[Update] 3rd place solution",
      "comment_count": 3,
      "views": 0,
      "votes": 22
    },
    {
      "id": 176405,
      "title": "8th place solution",
      "comment_count": 11,
      "views": 0,
      "votes": 31
    },
    {
      "id": 173572,
      "title": "Training tf records available here for download !!!",
      "comment_count": 12,
      "views": 0,
      "votes": 21
    },
    {
      "id": 175129,
      "title": "Some reverse engineering attempts",
      "comment_count": 2,
      "views": 0,
      "votes": 10
    },
    {
      "id": 175377,
      "title": "about baseline",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 174283,
      "title": "Why so many teams get map of 0.277? Is it a new baseline? ",
      "comment_count": 29,
      "views": 0,
      "votes": 9
    },
    {
      "id": 171795,
      "title": "Be early... Do nothing... Get points !!!",
      "comment_count": 20,
      "views": 0,
      "votes": 44
    },
    {
      "id": 174874,
      "title": "Competition Deadline Question",
      "comment_count": 8,
      "views": 0,
      "votes": 5
    },
    {
      "id": 175226,
      "title": "Recording your CV and LB, and also final score.",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 175404,
      "title": "Did anyone train on pytorch and convert it to tensorflow model?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 174856,
      "title": "how to train delf model with tpu?",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 175223,
      "title": "What happens if our submission is still processing when the competition ends?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 175224,
      "title": "test comment try",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 174824,
      "title": "What is going on with the avatars of the second team in LB?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 173262,
      "title": "Removing grayscale images boosts public LB score",
      "comment_count": 5,
      "views": 0,
      "votes": 12
    },
    {
      "id": 172566,
      "title": "Make This Competition More Accessible",
      "comment_count": 6,
      "views": 0,
      "votes": 32
    },
    {
      "id": 174076,
      "title": "About Scoring Script",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 174048,
      "title": "HELP uploading models creates tar files and not folders !",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 174478,
      "title": "How to download GLDv1",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 172414,
      "title": "Finetune the Baseline - LB 0.272+ ",
      "comment_count": 13,
      "views": 0,
      "votes": 13
    },
    {
      "id": 173881,
      "title": "Is L2 normalization important and why",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 170773,
      "title": "Exceeded Allowed Compute Error？model parameter size limit?GPU memory limit?",
      "comment_count": 17,
      "views": 0,
      "votes": 5
    },
    {
      "id": 170485,
      "title": "Submission is pending",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 172989,
      "title": "Are model trained offline legit ?",
      "comment_count": 4,
      "views": 0,
      "votes": 8
    },
    {
      "id": 173991,
      "title": "What's your local ROxford/RParis results?",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 173246,
      "title": "Anyone with tfrecords of Training Dataset ?",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 172706,
      "title": "Apart from Baseline Simple EfficientNet B0 (1 Epoch) 0.089",
      "comment_count": 3,
      "views": 0,
      "votes": 13
    },
    {
      "id": 173247,
      "title": "Anyone with tfrecords of Training Dataset ?",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 172471,
      "title": "Baseline model",
      "comment_count": 8,
      "views": 0,
      "votes": 8
    },
    {
      "id": 165685,
      "title": "Learning objective",
      "comment_count": 12,
      "views": 0,
      "votes": 2
    },
    {
      "id": 171297,
      "title": "Download tfrecords for training on delf here!!!",
      "comment_count": 1,
      "views": 0,
      "votes": 13
    },
    {
      "id": 173298,
      "title": "Run Time Requirement Question",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 173250,
      "title": "Submission Scoring Error",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 171916,
      "title": "TF Records Training data",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 172205,
      "title": "It all depends on how the triplets are selected !!!!",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 170477,
      "title": "Understanding the Competition Metric",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 172107,
      "title": "Will the competition length be renewed?",
      "comment_count": 0,
      "views": 0,
      "votes": -1
    },
    {
      "id": 171675,
      "title": "please explain me the role of partition in scoring script! ",
      "comment_count": 6,
      "views": 0,
      "votes": 1
    },
    {
      "id": 170363,
      "title": "Maybe relevant",
      "comment_count": 2,
      "views": 0,
      "votes": 22
    },
    {
      "id": 166354,
      "title": "How a large-scale image retrieval system works?",
      "comment_count": 1,
      "views": 0,
      "votes": 7
    },
    {
      "id": 171176,
      "title": "Internet connection and Pretrained weights",
      "comment_count": 3,
      "views": 0,
      "votes": 5
    },
    {
      "id": 170214,
      "title": "Alternative to Softmax?",
      "comment_count": 7,
      "views": 0,
      "votes": 5
    },
    {
      "id": 171476,
      "title": "Kaggle Infrastructure",
      "comment_count": 2,
      "views": 0,
      "votes": 0
    },
    {
      "id": 166149,
      "title": "Re-saving baseline model",
      "comment_count": 4,
      "views": 0,
      "votes": 13
    },
    {
      "id": 171256,
      "title": "Conditional Image Retrieval (CIR) ....(Maybe relevant)",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 166353,
      "title": "How are people without multiple GPUs supposed to participate?",
      "comment_count": 14,
      "views": 0,
      "votes": 9
    },
    {
      "id": 170523,
      "title": "CVPR 2020 Workshop - Image Matching local features & beyond",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 170993,
      "title": "Siamese Neural Network",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 163139,
      "title": "onnx -> tf",
      "comment_count": 10,
      "views": 0,
      "votes": 11
    },
    {
      "id": 169969,
      "title": "Please Report these Profiles.",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 166468,
      "title": "Are we allowed to train on our own machines?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 170768,
      "title": "Opinion poll",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 169369,
      "title": "i need a team",
      "comment_count": 7,
      "views": 0,
      "votes": 2
    },
    {
      "id": 170449,
      "title": "Get shape of TensorSpec",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 166518,
      "title": "When is the Recognition Challenge going to be released?",
      "comment_count": 1,
      "views": 0,
      "votes": 8
    },
    {
      "id": 170014,
      "title": "How to access kaggle commit output..???",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 169296,
      "title": "Public/Private Split",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 169304,
      "title": "Is GLD v1 allowed to train?",
      "comment_count": 2,
      "views": 0,
      "votes": null
    },
    {
      "id": 168868,
      "title": "Do I need to turn on the GPU?  ",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 168439,
      "title": "How should TFRecords be formatted?",
      "comment_count": 12,
      "views": 0,
      "votes": 0
    },
    {
      "id": 169154,
      "title": "Is it possible to do transfer learning with DELF model on tfhub?",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 168819,
      "title": "triplet loss doesnt converge",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 167512,
      "title": "Can we use Index data in our training?",
      "comment_count": 2,
      "views": 0,
      "votes": 3
    },
    {
      "id": 163287,
      "title": "Previous winning solutions ",
      "comment_count": 6,
      "views": 0,
      "votes": 47
    },
    {
      "id": 166913,
      "title": "why are so few teams participating?",
      "comment_count": 8,
      "views": 0,
      "votes": 9
    },
    {
      "id": 166335,
      "title": "Query regarding dataset",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 168004,
      "title": "How would you speed up convergence?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 167598,
      "title": "Not For Beginner",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 165386,
      "title": "Some questions about image preprocessing",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 167877,
      "title": "GCS Path Not Found Error",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 167602,
      "title": "hello kaggle worker, my notebook  takes too much time and it cancel automatically submit",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 167563,
      "title": "Imbalanced datasets",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 167529,
      "title": "Test data",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 165643,
      "title": "Some Questions about Understanding Data",
      "comment_count": 8,
      "views": 0,
      "votes": 3
    },
    {
      "id": 163799,
      "title": "Where to start?",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 167358,
      "title": "Submission Errors",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 164192,
      "title": "Training Image Resized Set",
      "comment_count": 2,
      "views": 0,
      "votes": 8
    },
    {
      "id": 167188,
      "title": "Datasets for judging a model",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 165574,
      "title": "Evaluation metric",
      "comment_count": 0,
      "views": 0,
      "votes": 25
    },
    {
      "id": 166422,
      "title": "🔴Kaggle Meetup:Landmark & Retrieval Landmark Recognition[Youtube Video]",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 166884,
      "title": "Notebook Threw Exception: Not able to figure out",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 166887,
      "title": "Submission Code Snippet",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 166850,
      "title": "How to do postprocessing?",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 166641,
      "title": "Are the Images in RGB or BGR?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 166808,
      "title": "Question on time limitation",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 164803,
      "title": "is the baseline  using Delf descriptors?",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 164647,
      "title": "First, second and third place solutions of Google Landmark Retrieval 2019",
      "comment_count": 2,
      "views": 0,
      "votes": 9
    },
    {
      "id": 166615,
      "title": "Can some share training details of first Baseline?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 164428,
      "title": "Submission failed. ",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 165092,
      "title": "Guide on how to submit Keras model",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 163086,
      "title": "Why is tensforflow required to submit a model ?",
      "comment_count": 16,
      "views": 0,
      "votes": 21
    },
    {
      "id": 165394,
      "title": "Notebook Exceeded Allowed Compute",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 163285,
      "title": "Why do you hate postprocessing, local features & RANSAC?",
      "comment_count": 10,
      "views": 0,
      "votes": 18
    },
    {
      "id": 163589,
      "title": "How to do data preprocessing while inference?",
      "comment_count": 13,
      "views": 0,
      "votes": 11
    },
    {
      "id": 165141,
      "title": "Number of Teams in the Leaderboard ",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 165246,
      "title": "Literature on ranking?",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 164321,
      "title": "Be ware - Train vs Test data statistics",
      "comment_count": 1,
      "views": 0,
      "votes": 7
    },
    {
      "id": 165120,
      "title": "Team Up to share the cost of GCP",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 163508,
      "title": "What are the constrains for generating embedding",
      "comment_count": 6,
      "views": 0,
      "votes": 4
    },
    {
      "id": 164004,
      "title": "An Improved Faster R-CNN for Same Object Retrieval - Relevant Research Paper",
      "comment_count": 2,
      "views": 0,
      "votes": 4
    },
    {
      "id": 164817,
      "title": "What is Media contributed by ETH?",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 164769,
      "title": "Summary of GLDv2 paper - 'A Large-scale benchmark for instance-level recognition and retrieval'",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 164757,
      "title": "How to use pytorch in this competiton?",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 163828,
      "title": "Important Research Papers for Competition",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 164676,
      "title": "too big train data",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 163837,
      "title": "1st Place Solution to Google Lankmark Retrieval 2019 ",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 164229,
      "title": "Team Thread",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 163390,
      "title": "Identifying landmark name from landmark id",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 164049,
      "title": "Ensemble Learning",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 163830,
      "title": "Notable Blogs for Competition",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 163175,
      "title": "\"index/\" ?",
      "comment_count": 3,
      "views": 0,
      "votes": 5
    },
    {
      "id": 163654,
      "title": "Embedding?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 163197,
      "title": "How big is the input size?",
      "comment_count": 4,
      "views": 0,
      "votes": null
    },
    {
      "id": 163290,
      "title": "Dataloader (WIP)",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    }
  ],
  "errors": []
}