{
  "total": 144,
  "topics": [
    {
      "id": 587927,
      "title": "Recap of Competition - Congratulations to the Winners!",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 572329,
      "title": "All you need know about Full-waveform Inversion",
      "comment_count": 31,
      "views": 0,
      "votes": 77
    },
    {
      "id": 572327,
      "title": "Welcome to the Yale/UNC-Chapel Hill – Geophysical Waveform Inversion competition!",
      "comment_count": 4,
      "views": 0,
      "votes": 15
    },
    {
      "id": 572279,
      "title": "Get Started Here!",
      "comment_count": 0,
      "views": 0,
      "votes": 8
    },
    {
      "id": 572278,
      "title": "This Competition has an Official Discord Channel",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 587950,
      "title": "2nd place solution with code: refinement with FWI",
      "comment_count": 19,
      "views": 0,
      "votes": 49
    },
    {
      "id": 587344,
      "title": "Some cheating teams！！！",
      "comment_count": 30,
      "views": 0,
      "votes": 28
    },
    {
      "id": 586283,
      "title": "Will Gold Medal be Under Public LB=12.0?",
      "comment_count": 73,
      "views": 0,
      "votes": 25
    },
    {
      "id": 587915,
      "title": "69th place solution",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 587388,
      "title": "1st Place Solution",
      "comment_count": 66,
      "views": 0,
      "votes": 194
    },
    {
      "id": 587443,
      "title": "Team FAMAS. 5th place solution.",
      "comment_count": 15,
      "views": 0,
      "votes": 43
    },
    {
      "id": 587419,
      "title": "3rd place solution",
      "comment_count": 15,
      "views": 0,
      "votes": 60
    },
    {
      "id": 587460,
      "title": "6th Place Solution Summary",
      "comment_count": 11,
      "views": 0,
      "votes": 45
    },
    {
      "id": 587500,
      "title": "4th place solution for the GWI competition  ",
      "comment_count": 9,
      "views": 0,
      "votes": 42
    },
    {
      "id": 587498,
      "title": "9th place - Custom CUDA kernel for wave propagation",
      "comment_count": 10,
      "views": 0,
      "votes": 59
    },
    {
      "id": 587511,
      "title": "15th place: Two-stage model",
      "comment_count": 12,
      "views": 0,
      "votes": 30
    },
    {
      "id": 587553,
      "title": "45th Place Solution",
      "comment_count": 4,
      "views": 0,
      "votes": 11
    },
    {
      "id": 587529,
      "title": "14th place solution",
      "comment_count": 5,
      "views": 0,
      "votes": 24
    },
    {
      "id": 587402,
      "title": "20th solution - Caformer + data generation",
      "comment_count": 2,
      "views": 0,
      "votes": 46
    },
    {
      "id": 587506,
      "title": "Tips to make forward model efficient",
      "comment_count": 25,
      "views": 0,
      "votes": 6
    },
    {
      "id": 587389,
      "title": "How far did you get by continuing to train Bartley's type models?",
      "comment_count": 26,
      "views": 0,
      "votes": 9
    },
    {
      "id": 587412,
      "title": "10th Place Solution (monnu part)",
      "comment_count": 0,
      "views": 0,
      "votes": 34
    },
    {
      "id": 587404,
      "title": "12th Solution",
      "comment_count": 4,
      "views": 0,
      "votes": 31
    },
    {
      "id": 587429,
      "title": "10th Place Solution (yu4u's part)",
      "comment_count": 2,
      "views": 0,
      "votes": 25
    },
    {
      "id": 587390,
      "title": "24th Place Silver Medal Solution",
      "comment_count": 4,
      "views": 0,
      "votes": 17
    },
    {
      "id": 587386,
      "title": "[Our findings] Did you read about how real seismic survey work when there is no ground truth of subsurface?",
      "comment_count": 14,
      "views": 0,
      "votes": 11
    },
    {
      "id": 587387,
      "title": "this is first time in kaggle history?",
      "comment_count": 9,
      "views": 0,
      "votes": 9
    },
    {
      "id": 587457,
      "title": "How much compute did you use?",
      "comment_count": 11,
      "views": 0,
      "votes": 6
    },
    {
      "id": 587766,
      "title": "Cheap post-processing to fight on medal borders",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 587417,
      "title": "Share some findings",
      "comment_count": 1,
      "views": 0,
      "votes": 8
    },
    {
      "id": 587494,
      "title": "63th place solution —— Overview",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 587841,
      "title": "The solution ranked 68th",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 587497,
      "title": "Cheating is a serious problem, not only in Geophysical Waveform Inversion, but also in other competitions.",
      "comment_count": 0,
      "views": 0,
      "votes": 5
    },
    {
      "id": 586264,
      "title": "I did it again: submit twice the same sub",
      "comment_count": 42,
      "views": 0,
      "votes": 13
    },
    {
      "id": 587406,
      "title": "Was heavy GPU computation required to achieve a medal in this competition?",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 587422,
      "title": " Some Findings on Backpropagation with Forward Models",
      "comment_count": 4,
      "views": 0,
      "votes": 3
    },
    {
      "id": 587434,
      "title": "adjoint method of computing velocity gradient for wave modeling",
      "comment_count": 8,
      "views": 0,
      "votes": null
    },
    {
      "id": 587395,
      "title": "102nd Place Solution - Clustering And Refinement with Reconstruction Loss To Beat Public Codes",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 586363,
      "title": "Will winner score be above or below 5?",
      "comment_count": 26,
      "views": 0,
      "votes": 8
    },
    {
      "id": 587450,
      "title": "OpenFWI UINT8+1 quantized dataset",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 587636,
      "title": "Just noticed something special",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 585815,
      "title": "Sharing a few ideas (+ torch vel_to_seis)",
      "comment_count": 7,
      "views": 0,
      "votes": 8
    },
    {
      "id": 586575,
      "title": "Divergence of the val_loss",
      "comment_count": 5,
      "views": 0,
      "votes": 3
    },
    {
      "id": 585450,
      "title": "STELLAR performance from Harshit Sheoran",
      "comment_count": 6,
      "views": 0,
      "votes": 9
    },
    {
      "id": 586099,
      "title": "Coding Competition ",
      "comment_count": 7,
      "views": 0,
      "votes": null
    },
    {
      "id": 585128,
      "title": "What's your guess of the number of teams that will break the 15.0 barrier?",
      "comment_count": 39,
      "views": 0,
      "votes": 15
    },
    {
      "id": 586109,
      "title": "A question about nn.DataParallel",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 585456,
      "title": "Some do great, others don't !",
      "comment_count": 14,
      "views": 0,
      "votes": 26
    },
    {
      "id": 585273,
      "title": "He jumped from 25.1 to 16.1👏",
      "comment_count": 30,
      "views": 0,
      "votes": 20
    },
    {
      "id": 585539,
      "title": "[Let's Collaborate] Generating `CurveVel` & `CurveFault` for Augmentation",
      "comment_count": 12,
      "views": 0,
      "votes": 8
    },
    {
      "id": 587157,
      "title": "# Kaggle FWI Competition Notebook",
      "comment_count": 0,
      "views": 0,
      "votes": -15
    },
    {
      "id": 582801,
      "title": "A Better Way To Ensemble",
      "comment_count": 11,
      "views": 0,
      "votes": 51
    },
    {
      "id": 585236,
      "title": "CUDA Speed Up for vel-to-seis and Some Findings",
      "comment_count": 17,
      "views": 0,
      "votes": 14
    },
    {
      "id": 585694,
      "title": "New paper on Waveform inversion",
      "comment_count": 6,
      "views": 0,
      "votes": 3
    },
    {
      "id": 586103,
      "title": "TPU issues with transformers library",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 585194,
      "title": "Anyteam can make Swin Transformer work?",
      "comment_count": 8,
      "views": 0,
      "votes": 11
    },
    {
      "id": 584426,
      "title": "anyone try to use Physics Driven Neural Network for Interval",
      "comment_count": 4,
      "views": 0,
      "votes": -1
    },
    {
      "id": 585602,
      "title": "Best current beginner notebook that runs here on kaggle ?",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 584772,
      "title": "my maybe efficient trainer?",
      "comment_count": 39,
      "views": 0,
      "votes": 16
    },
    {
      "id": 583217,
      "title": "Beware of Bartley!",
      "comment_count": 18,
      "views": 0,
      "votes": 43
    },
    {
      "id": 585609,
      "title": "Assessing Model Convergence: Should I Continue Training?",
      "comment_count": 1,
      "views": 0,
      "votes": 3
    },
    {
      "id": 585834,
      "title": "dm if anyone needs compute",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 585689,
      "title": "How is everyone dealing with the size of the data?",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 583896,
      "title": "Tips to speed up training",
      "comment_count": 33,
      "views": 0,
      "votes": 66
    },
    {
      "id": 585255,
      "title": "what’s the lowest CV/LB a single model can achieve?",
      "comment_count": 16,
      "views": 0,
      "votes": 3
    },
    {
      "id": 585228,
      "title": "Subs not working? ",
      "comment_count": 5,
      "views": 0,
      "votes": 1
    },
    {
      "id": 584794,
      "title": "MAE LB results questions. the LB results is 1000*MAE?",
      "comment_count": 2,
      "views": 0,
      "votes": -1
    },
    {
      "id": 583016,
      "title": "LR Decay For MAE",
      "comment_count": 5,
      "views": 0,
      "votes": 39
    },
    {
      "id": 584978,
      "title": "TeamUp - Best CV for dataset wise [ 7 days left ]",
      "comment_count": 3,
      "views": 0,
      "votes": 4
    },
    {
      "id": 584640,
      "title": "Additional data, efwi dataset",
      "comment_count": 5,
      "views": 0,
      "votes": 12
    },
    {
      "id": 584739,
      "title": "Search team",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 582785,
      "title": "CAFormer Improved - [CV 24.2 LB 28.8]",
      "comment_count": 70,
      "views": 0,
      "votes": 78
    },
    {
      "id": 584959,
      "title": "25.6 LB leak",
      "comment_count": 11,
      "views": 0,
      "votes": -7
    },
    {
      "id": 584580,
      "title": "How to break Bartley's strong base LB 28.8? ",
      "comment_count": 11,
      "views": 0,
      "votes": 2
    },
    {
      "id": 584560,
      "title": "How should I split data for cross-validation with 940 files from 10 distinct families?",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 584008,
      "title": "Some statistics related info of the OpenFWI dataset.",
      "comment_count": 3,
      "views": 0,
      "votes": 4
    },
    {
      "id": 583872,
      "title": "Epoch Train/CV scores of all datasets [Can we break Bartley's strong models]",
      "comment_count": 9,
      "views": 0,
      "votes": 6
    },
    {
      "id": 584137,
      "title": "CAFormer 8Core TPU Training(Pytorch/XLA)",
      "comment_count": 7,
      "views": 0,
      "votes": 7
    },
    {
      "id": 584747,
      "title": "Which model is best for you",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 583763,
      "title": "Would someone work with me by sharing GPU?",
      "comment_count": 10,
      "views": 0,
      "votes": 3
    },
    {
      "id": 583927,
      "title": "Has someone managed to generate correct seismic data using Deepwave?",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 583866,
      "title": "Why leaderboard score is worse than cross-validation",
      "comment_count": 7,
      "views": 0,
      "votes": 7
    },
    {
      "id": 583946,
      "title": "About the Leaderboard",
      "comment_count": 14,
      "views": 0,
      "votes": 1
    },
    {
      "id": 584142,
      "title": "Acoustic Wave Equation Forward Modeling ",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 583908,
      "title": "Are the velocities always between 1500 m/s and 4500 m/s?",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 583815,
      "title": "[Convent Train Failed] Has anyone else encountered the same issue?",
      "comment_count": 18,
      "views": 0,
      "votes": 2
    },
    {
      "id": 580689,
      "title": "Augmentation Ideas?",
      "comment_count": 17,
      "views": 0,
      "votes": 18
    },
    {
      "id": 582845,
      "title": "A Two-Stage Approach for Prediction using CNN Classification and InversionNet[LB: 133.6]",
      "comment_count": 19,
      "views": 0,
      "votes": 12
    },
    {
      "id": 583385,
      "title": "Downloading a Specific File or Folder from a Kaggle Dataset via API",
      "comment_count": 7,
      "views": 0,
      "votes": 2
    },
    {
      "id": 583392,
      "title": "Why does fine-tuning make mae larger",
      "comment_count": 12,
      "views": 0,
      "votes": 2
    },
    {
      "id": 583343,
      "title": "Complete openFWI Dataset",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 583453,
      "title": "How are the input and output data fundamentally related?",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 583491,
      "title": "Do we know the position of the receivers? Is it one-dimensional (1D) or two-dimensional (2D)?",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 583538,
      "title": "How to use LoRA on timm？peft？",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 583715,
      "title": "Conflict time steps between paper and data",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 580892,
      "title": "Best single CurveFault_B model?",
      "comment_count": 7,
      "views": 0,
      "votes": 5
    },
    {
      "id": 583563,
      "title": "Physics informed Loss?",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 579376,
      "title": "unsupervised training with test Waveform?",
      "comment_count": 14,
      "views": 0,
      "votes": 19
    },
    {
      "id": 583407,
      "title": "Is it acceptable to create a dataset based on the test set and use it for the submission?",
      "comment_count": 4,
      "views": 0,
      "votes": 1
    },
    {
      "id": 583375,
      "title": "Is the training code based on Egor Trushin's float16 UNet notebook?",
      "comment_count": 0,
      "views": 0,
      "votes": 1
    },
    {
      "id": 573571,
      "title": "Attempt to create OpenFWI seis data",
      "comment_count": 21,
      "views": 0,
      "votes": 48
    },
    {
      "id": 582622,
      "title": "Physics-Driven Neural Network for Interval Q Inversion",
      "comment_count": 4,
      "views": 0,
      "votes": 2
    },
    {
      "id": 572747,
      "title": "A Clear Explanation of The Dataset",
      "comment_count": 36,
      "views": 0,
      "votes": 156
    },
    {
      "id": 575846,
      "title": "Which columns to drop for submission?",
      "comment_count": 5,
      "views": 0,
      "votes": 5
    },
    {
      "id": 579841,
      "title": "ConvNeXt Approach - [CV 31.9 LB 36.4]",
      "comment_count": 53,
      "views": 0,
      "votes": 84
    },
    {
      "id": 582597,
      "title": "Full resolution data vs compressed data performance difference?",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 581134,
      "title": "How to use TPU in pytorch rightly?",
      "comment_count": 11,
      "views": 0,
      "votes": 6
    },
    {
      "id": 580988,
      "title": "There is a leak in train files",
      "comment_count": 6,
      "views": 0,
      "votes": 30
    },
    {
      "id": 575917,
      "title": "You can get more GPU 45h-60h per week on Kaggle with Colab Pro",
      "comment_count": 6,
      "views": 0,
      "votes": 9
    },
    {
      "id": 581298,
      "title": "which session",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 581204,
      "title": "Execute Time Restriction",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 580980,
      "title": "Failed to load the model",
      "comment_count": 8,
      "views": 0,
      "votes": 1
    },
    {
      "id": 575972,
      "title": "CV vs LB Scores [Check Bartley's Awesome Posts/Notebooks/Comments - better CV/LB]",
      "comment_count": 9,
      "views": 0,
      "votes": 9
    },
    {
      "id": 572867,
      "title": "Questions on placement of sources and sensors.",
      "comment_count": 15,
      "views": 0,
      "votes": 13
    },
    {
      "id": 579965,
      "title": "Seismic source function",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 579727,
      "title": "Performance of trained models solely based on the competition data",
      "comment_count": 3,
      "views": 0,
      "votes": 1
    },
    {
      "id": 578305,
      "title": "HGNet-V2 Encoder - [CV 55.6 LB 60.9]",
      "comment_count": 22,
      "views": 0,
      "votes": 45
    },
    {
      "id": 579040,
      "title": "The Power of Auxiliary Loss Functions in Minimizing MAE Loss",
      "comment_count": 2,
      "views": 0,
      "votes": 16
    },
    {
      "id": 572434,
      "title": "Huggingface + Kaggle OpenFWI Dataset 670G available + Training",
      "comment_count": 26,
      "views": 0,
      "votes": 53
    },
    {
      "id": 578102,
      "title": "Fva_l1 model",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 576237,
      "title": "Question on Scaling and Offset from best public notebooks ",
      "comment_count": 13,
      "views": 0,
      "votes": 9
    },
    {
      "id": 572324,
      "title": "Having problem downloading data? [solved]",
      "comment_count": 8,
      "views": 0,
      "votes": 5
    },
    {
      "id": 575097,
      "title": "Improved UNet pipepline with larger dataset",
      "comment_count": 12,
      "views": 0,
      "votes": 32
    },
    {
      "id": 572314,
      "title": "Starting materials and references",
      "comment_count": 1,
      "views": 0,
      "votes": 13
    },
    {
      "id": 575243,
      "title": "is classification + image2image? individual model or single model for 10 datasets?",
      "comment_count": 20,
      "views": 0,
      "votes": 18
    },
    {
      "id": 576243,
      "title": "Clarification on Source/Receiver Line Alignment and Velocity Map Orientation",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 575178,
      "title": "A new SOTA (?) paper \"A Unified Framework for Forward and Inverse Problems in Subsurface Imaging Using Latent Space Translations.\"",
      "comment_count": 5,
      "views": 0,
      "votes": 15
    },
    {
      "id": 575865,
      "title": "Can not download OpenFWI",
      "comment_count": 2,
      "views": 0,
      "votes": 1
    },
    {
      "id": 575166,
      "title": "InversionNet using \"CRFs\" ",
      "comment_count": 5,
      "views": 0,
      "votes": 2
    },
    {
      "id": 574648,
      "title": "model diverges",
      "comment_count": 9,
      "views": 0,
      "votes": 9
    },
    {
      "id": 574495,
      "title": "Scaling up results- Kaggle data vs full data",
      "comment_count": 8,
      "views": 0,
      "votes": 27
    },
    {
      "id": 572399,
      "title": "[LB 249.7] OpenFWI Models  -- [LB  130.5] Ensemble",
      "comment_count": 5,
      "views": 0,
      "votes": 11
    },
    {
      "id": 574083,
      "title": "Here's Why They Call It a ‘Velocity’ Model When Nothing’s Actually Moving.",
      "comment_count": 2,
      "views": 0,
      "votes": 7
    },
    {
      "id": 574284,
      "title": "Question about submission file format",
      "comment_count": 2,
      "views": 0,
      "votes": 2
    },
    {
      "id": 573319,
      "title": "ML vs non-ML solutions — fairness of the competition",
      "comment_count": 5,
      "views": 0,
      "votes": 7
    },
    {
      "id": 572910,
      "title": "No machine learning needed!",
      "comment_count": 9,
      "views": 0,
      "votes": -22
    },
    {
      "id": 572506,
      "title": "No support of Firefox",
      "comment_count": 7,
      "views": 0,
      "votes": -2
    },
    {
      "id": 572552,
      "title": "Not the norm",
      "comment_count": 1,
      "views": 0,
      "votes": 1
    },
    {
      "id": 572334,
      "title": "is all OpenFWI Dataset ~ 670G vs Kaggle Dataset ~ Train 14TB (dataset) / Pre-trained Models(dataset)",
      "comment_count": 7,
      "views": 0,
      "votes": 21
    },
    {
      "id": 572796,
      "title": "11th on the leader board - CNNs work ?",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 572365,
      "title": "📶🌊Good starter resources for Yale/UNC GWI contest 📶🌊",
      "comment_count": 0,
      "views": 0,
      "votes": 10
    },
    {
      "id": 572330,
      "title": "[Basic Geophysics] Full Waveform Inversion",
      "comment_count": 0,
      "views": 0,
      "votes": 9
    },
    {
      "id": 572638,
      "title": "Competition ",
      "comment_count": 1,
      "views": 0,
      "votes": null
    },
    {
      "id": 572561,
      "title": "Model and Data details",
      "comment_count": 1,
      "views": 0,
      "votes": null
    }
  ],
  "errors": []
}