{
  "id": 306246,
  "title": "Resources for fast start! ",
  "url": "/competitions/happy-whale-and-dolphin/discussion/306246",
  "author_name": "",
  "post_date": "2022-02-08T18:19:59.291736700Z",
  "votes": 82,
  "comment_count": 12,
  "views": 0,
  "content": "<h1>Exploratory Data Analysis</h1>\n<p>At beginning of the competition, you should do Exploratory Data Analysis, to find some relations in features and targets, write down some ideas for Feature Engineering from these explorations/insights, and much other stuff.</p>\n<ul>\n<li><a href=\"https://www.kaggle.com/awsaf49/happywhale-data-distribution/\" target=\"_blank\">Happywhale: Data Distribution 🐋🐬</a> by <a href=\"https://www.kaggle.com/awsaf49\" target=\"_blank\">@awsaf49</a>.</li>\n<li><a href=\"https://www.kaggle.com/gpreda/happy-whales-and-dolphins#Images-data-exploration\" target=\"_blank\">Happy Whales and Dolphins</a> by <a href=\"https://www.kaggle.com/gpreda\" target=\"_blank\">@gpreda</a>.</li>\n<li><a href=\"https://www.kaggle.com/ruchi798/and-identification-eda-augmentation\" target=\"_blank\">🐋 and 🐬 Identification: EDA + Augmentation</a> by <a href=\"https://www.kaggle.com/ruchi798\" target=\"_blank\">@ruchi798</a>.</li>\n<li><a href=\"https://www.kaggle.com/abhranta/starter-eda-aug\" target=\"_blank\">🐋🐬 STARTER+EDA+AUG</a> by <a href=\"https://www.kaggle.com/abhranta\" target=\"_blank\">@abhranta</a>.</li>\n<li><a href=\"https://www.kaggle.com/bsridatta/happywhale\" target=\"_blank\">HappyWhale 🐳</a> by <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>.</li>\n<li><a href=\"https://www.kaggle.com/vad13irt/happywhale-exploratory-data-analysis/notebook\" target=\"_blank\">Happywhale: Exploratory Data Analysis</a> by <strong>me</strong>.</li>\n<li><a href=\"https://www.kaggle.com/hamditarek/whale-and-dolphin-identification-eda\" target=\"_blank\">Whale and Dolphin Identification EDA</a> by <a href=\"https://www.kaggle.com/hamditarek\" target=\"_blank\">@hamditarek</a>.</li>\n<li><a href=\"https://www.kaggle.com/ritesh2000/happywhale-eda-fastai-albumentations-starter\" target=\"_blank\">HappyWhale EDA,Fastai &amp; Albumentations Starter🦈🐳</a> by <a href=\"https://www.kaggle.com/ritesh2000\" target=\"_blank\">@ritesh2000</a>.</li>\n<li><a href=\"https://www.kaggle.com/dschettler8845/eda-and-baseline-solution\" target=\"_blank\">😊🐳&amp;🐬 - EDA and Baseline Solution</a> by <a href=\"https://www.kaggle.com/dschettler8845\" target=\"_blank\">@dschettler8845</a>.</li>\n</ul>\n<h1>Metric/Loss Understanding</h1>\n<p>The intermediate step, which also is one of the most important, is Metric Understanding. Choosing the right metric is crucial while evaluating machine learning (ML) models. Various metrics are proposed to evaluate ML models in different applications, and I thought it may be helpful to provide a summary of popular metrics here, for a better understanding of each metric and the applications they can be used for. In some applications looking at a single metric may not give you the whole picture of the problem you are solving, and you may want to use a subset of the metrics discussed in this post to have a concrete evaluation of your models. </p>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304615\" target=\"_blank\">The MAP@5 metric</a> by <a href=\"https://www.kaggle.com/kishalmandal\" target=\"_blank\">@kishalmandal</a>.</li>\n</ul>\n<h1>Modeling</h1>\n<p>The most important step is to build a good model, that can well predict unseen data (aka Test Dataset), so choosing the right model is often not easy, it requires a lot of experiments and of course experience.</p>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304973\" target=\"_blank\">Papers on Whale and Dolphin Identification 🐬🐋</a> by <a href=\"https://www.kaggle.com/datascientistfp\" target=\"_blank\">@datascientistfp</a>.</li>\n<li><a href=\"https://www.kaggle.com/ks2019/happywhale-arcface-baseline-tpu\" target=\"_blank\">HappyWhale ArcFace Baseline (TPU)</a> by <a href=\"https://www.kaggle.com/ks2019\" target=\"_blank\">@ks2019</a>.</li>\n<li><a href=\"https://www.kaggle.com/debarshichanda/pytorch-happywhale-siamese-starter\" target=\"_blank\">[Pytorch] HappyWhale Siamese Starter</a> by <a href=\"https://www.kaggle.com/debarshichanda\" target=\"_blank\">@debarshichanda</a>.</li>\n<li><a href=\"https://www.kaggle.com/debarshichanda/pytorch-arcface-gem-pooling-starter\" target=\"_blank\">[Pytorch] ArcFace + GeM Pooling Starter</a> by <a href=\"https://www.kaggle.com/debarshichanda\" target=\"_blank\">@debarshichanda</a>.</li>\n<li><a href=\"https://www.kaggle.com/ammarnassanalhajali/cnn-with-keras-stater\" target=\"_blank\">CNN with Keras Stater</a> by <a href=\"https://www.kaggle.com/ammarnassanalhajali\" target=\"_blank\">@ammarnassanalhajali</a>.</li>\n<li><a href=\"https://www.kaggle.com/ayuraj/train-pytorch-finetune-convnext\" target=\"_blank\">[Train][PyTorch] FineTune ConvNeXt</a> by <a href=\"https://www.kaggle.com/ayuraj\" target=\"_blank\">@ayuraj</a>.</li>\n<li><a href=\"https://www.kaggle.com/meowmeowmeowmeowmeow/species-classification-starter-tf-data-pipeline\" target=\"_blank\">Species classification starter + TF data pipeline</a> by <a href=\"https://www.kaggle.com/meowmeowmeowmeowmeow\" target=\"_blank\">@meowmeowmeowmeowmeow</a>.</li>\n<li><a href=\"https://www.kaggle.com/sentrankim/sota2020-curricularface-amp-mixup-cutmix-sam\" target=\"_blank\">SOTA2020 CurricularFace+AMP+Mixup+CutMix+SAM</a> by <a href=\"https://www.kaggle.com/sentrankim\" target=\"_blank\">@sentrankim</a>.</li>\n</ul>\n<h1>Resting</h1>\n<p>Taking some rest for no long time is very important so you keep your brain fresh and avoid burnout. Also, it is useful to spend this period of time with the benefit of learning some additional materials for certain competitions.</p>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305919\" target=\"_blank\">What do the species look like?</a> by <a href=\"https://www.kaggle.com/andradaolteanu\" target=\"_blank\">@andradaolteanu</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304509\" target=\"_blank\">The most famous Whales!</a> by <strong>me</strong>.</li>\n</ul>\n<h1>Intersting insights</h1>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304633\" target=\"_blank\">Duplicate names in species, can be merged together</a> by <a href=\"https://www.kaggle.com/karthickp6\" target=\"_blank\">@karthickp6</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305428\" target=\"_blank\">LB probing and train/test split</a> by <a href=\"https://www.kaggle.com/olegsidorshin\" target=\"_blank\">@olegsidorshin</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305607\" target=\"_blank\">Fin detect, extract and identify pipeline with paper and code</a> by <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305843\" target=\"_blank\">Avoid flipping the images!</a> by <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305030\" target=\"_blank\">Additional Metadata For Train/Test [Path and Image Size]</a> by <a href=\"https://www.kaggle.com/dschettler8845\" target=\"_blank\">@dschettler8845</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305899\" target=\"_blank\">Some images have more than one whale/Dolphin</a> by <a href=\"https://www.kaggle.com/chihantsai\" target=\"_blank\">@chihantsai</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304536\" target=\"_blank\">Was Train Dataset applied Augmentations?</a> by <strong>me</strong>.</li>\n</ul>",
  "messages": [
    {
      "id": "1681821",
      "postDate": "02/08/2022 18:19:59",
      "content": "<h1>Exploratory Data Analysis</h1>\n<p>At beginning of the competition, you should do Exploratory Data Analysis, to find some relations in features and targets, write down some ideas for Feature Engineering from these explorations/insights, and much other stuff.</p>\n<ul>\n<li><a href=\"https://www.kaggle.com/awsaf49/happywhale-data-distribution/\" target=\"_blank\">Happywhale: Data Distribution 🐋🐬</a> by <a href=\"https://www.kaggle.com/awsaf49\" target=\"_blank\">@awsaf49</a>.</li>\n<li><a href=\"https://www.kaggle.com/gpreda/happy-whales-and-dolphins#Images-data-exploration\" target=\"_blank\">Happy Whales and Dolphins</a> by <a href=\"https://www.kaggle.com/gpreda\" target=\"_blank\">@gpreda</a>.</li>\n<li><a href=\"https://www.kaggle.com/ruchi798/and-identification-eda-augmentation\" target=\"_blank\">🐋 and 🐬 Identification: EDA + Augmentation</a> by <a href=\"https://www.kaggle.com/ruchi798\" target=\"_blank\">@ruchi798</a>.</li>\n<li><a href=\"https://www.kaggle.com/abhranta/starter-eda-aug\" target=\"_blank\">🐋🐬 STARTER+EDA+AUG</a> by <a href=\"https://www.kaggle.com/abhranta\" target=\"_blank\">@abhranta</a>.</li>\n<li><a href=\"https://www.kaggle.com/bsridatta/happywhale\" target=\"_blank\">HappyWhale 🐳</a> by <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>.</li>\n<li><a href=\"https://www.kaggle.com/vad13irt/happywhale-exploratory-data-analysis/notebook\" target=\"_blank\">Happywhale: Exploratory Data Analysis</a> by <strong>me</strong>.</li>\n<li><a href=\"https://www.kaggle.com/hamditarek/whale-and-dolphin-identification-eda\" target=\"_blank\">Whale and Dolphin Identification EDA</a> by <a href=\"https://www.kaggle.com/hamditarek\" target=\"_blank\">@hamditarek</a>.</li>\n<li><a href=\"https://www.kaggle.com/ritesh2000/happywhale-eda-fastai-albumentations-starter\" target=\"_blank\">HappyWhale EDA,Fastai &amp; Albumentations Starter🦈🐳</a> by <a href=\"https://www.kaggle.com/ritesh2000\" target=\"_blank\">@ritesh2000</a>.</li>\n<li><a href=\"https://www.kaggle.com/dschettler8845/eda-and-baseline-solution\" target=\"_blank\">😊🐳&amp;🐬 - EDA and Baseline Solution</a> by <a href=\"https://www.kaggle.com/dschettler8845\" target=\"_blank\">@dschettler8845</a>.</li>\n</ul>\n<h1>Metric/Loss Understanding</h1>\n<p>The intermediate step, which also is one of the most important, is Metric Understanding. Choosing the right metric is crucial while evaluating machine learning (ML) models. Various metrics are proposed to evaluate ML models in different applications, and I thought it may be helpful to provide a summary of popular metrics here, for a better understanding of each metric and the applications they can be used for. In some applications looking at a single metric may not give you the whole picture of the problem you are solving, and you may want to use a subset of the metrics discussed in this post to have a concrete evaluation of your models. </p>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304615\" target=\"_blank\">The MAP@5 metric</a> by <a href=\"https://www.kaggle.com/kishalmandal\" target=\"_blank\">@kishalmandal</a>.</li>\n</ul>\n<h1>Modeling</h1>\n<p>The most important step is to build a good model, that can well predict unseen data (aka Test Dataset), so choosing the right model is often not easy, it requires a lot of experiments and of course experience.</p>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304973\" target=\"_blank\">Papers on Whale and Dolphin Identification 🐬🐋</a> by <a href=\"https://www.kaggle.com/datascientistfp\" target=\"_blank\">@datascientistfp</a>.</li>\n<li><a href=\"https://www.kaggle.com/ks2019/happywhale-arcface-baseline-tpu\" target=\"_blank\">HappyWhale ArcFace Baseline (TPU)</a> by <a href=\"https://www.kaggle.com/ks2019\" target=\"_blank\">@ks2019</a>.</li>\n<li><a href=\"https://www.kaggle.com/debarshichanda/pytorch-happywhale-siamese-starter\" target=\"_blank\">[Pytorch] HappyWhale Siamese Starter</a> by <a href=\"https://www.kaggle.com/debarshichanda\" target=\"_blank\">@debarshichanda</a>.</li>\n<li><a href=\"https://www.kaggle.com/debarshichanda/pytorch-arcface-gem-pooling-starter\" target=\"_blank\">[Pytorch] ArcFace + GeM Pooling Starter</a> by <a href=\"https://www.kaggle.com/debarshichanda\" target=\"_blank\">@debarshichanda</a>.</li>\n<li><a href=\"https://www.kaggle.com/ammarnassanalhajali/cnn-with-keras-stater\" target=\"_blank\">CNN with Keras Stater</a> by <a href=\"https://www.kaggle.com/ammarnassanalhajali\" target=\"_blank\">@ammarnassanalhajali</a>.</li>\n<li><a href=\"https://www.kaggle.com/ayuraj/train-pytorch-finetune-convnext\" target=\"_blank\">[Train][PyTorch] FineTune ConvNeXt</a> by <a href=\"https://www.kaggle.com/ayuraj\" target=\"_blank\">@ayuraj</a>.</li>\n<li><a href=\"https://www.kaggle.com/meowmeowmeowmeowmeow/species-classification-starter-tf-data-pipeline\" target=\"_blank\">Species classification starter + TF data pipeline</a> by <a href=\"https://www.kaggle.com/meowmeowmeowmeowmeow\" target=\"_blank\">@meowmeowmeowmeowmeow</a>.</li>\n<li><a href=\"https://www.kaggle.com/sentrankim/sota2020-curricularface-amp-mixup-cutmix-sam\" target=\"_blank\">SOTA2020 CurricularFace+AMP+Mixup+CutMix+SAM</a> by <a href=\"https://www.kaggle.com/sentrankim\" target=\"_blank\">@sentrankim</a>.</li>\n</ul>\n<h1>Resting</h1>\n<p>Taking some rest for no long time is very important so you keep your brain fresh and avoid burnout. Also, it is useful to spend this period of time with the benefit of learning some additional materials for certain competitions.</p>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305919\" target=\"_blank\">What do the species look like?</a> by <a href=\"https://www.kaggle.com/andradaolteanu\" target=\"_blank\">@andradaolteanu</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304509\" target=\"_blank\">The most famous Whales!</a> by <strong>me</strong>.</li>\n</ul>\n<h1>Intersting insights</h1>\n<ul>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304633\" target=\"_blank\">Duplicate names in species, can be merged together</a> by <a href=\"https://www.kaggle.com/karthickp6\" target=\"_blank\">@karthickp6</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305428\" target=\"_blank\">LB probing and train/test split</a> by <a href=\"https://www.kaggle.com/olegsidorshin\" target=\"_blank\">@olegsidorshin</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305607\" target=\"_blank\">Fin detect, extract and identify pipeline with paper and code</a> by <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305843\" target=\"_blank\">Avoid flipping the images!</a> by <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305030\" target=\"_blank\">Additional Metadata For Train/Test [Path and Image Size]</a> by <a href=\"https://www.kaggle.com/dschettler8845\" target=\"_blank\">@dschettler8845</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305899\" target=\"_blank\">Some images have more than one whale/Dolphin</a> by <a href=\"https://www.kaggle.com/chihantsai\" target=\"_blank\">@chihantsai</a>.</li>\n<li><a href=\"https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304536\" target=\"_blank\">Was Train Dataset applied Augmentations?</a> by <strong>me</strong>.</li>\n</ul>",
      "rawMarkdown": "<h1>Exploratory Data Analysis</h1>\nAt beginning of the competition, you should do Exploratory Data Analysis, to find some relations in features and targets, write down some ideas for Feature Engineering from these explorations/insights, and much other stuff.\n\n\n- [Happywhale: Data Distribution 🐋🐬](https://www.kaggle.com/awsaf49/happywhale-data-distribution/) by @awsaf49.\n- [Happy Whales and Dolphins](https://www.kaggle.com/gpreda/happy-whales-and-dolphins#Images-data-exploration) by @gpreda.\n- [🐋 and 🐬 Identification: EDA + Augmentation](https://www.kaggle.com/ruchi798/and-identification-eda-augmentation) by @ruchi798.\n- [🐋🐬 STARTER+EDA+AUG](https://www.kaggle.com/abhranta/starter-eda-aug) by @abhranta.\n- [HappyWhale 🐳](https://www.kaggle.com/bsridatta/happywhale) by @bsridatta.\n- [Happywhale: Exploratory Data Analysis](https://www.kaggle.com/vad13irt/happywhale-exploratory-data-analysis/notebook) by **me**.\n- [Whale and Dolphin Identification EDA](https://www.kaggle.com/hamditarek/whale-and-dolphin-identification-eda) by @hamditarek.\n- [HappyWhale EDA,Fastai & Albumentations Starter🦈🐳](https://www.kaggle.com/ritesh2000/happywhale-eda-fastai-albumentations-starter) by @ritesh2000.\n- [😊🐳&🐬 - EDA and Baseline Solution](https://www.kaggle.com/dschettler8845/eda-and-baseline-solution) by @dschettler8845.\n\n\n<h1>Metric/Loss Understanding</h1>\nThe intermediate step, which also is one of the most important, is Metric Understanding. Choosing the right metric is crucial while evaluating machine learning (ML) models. Various metrics are proposed to evaluate ML models in different applications, and I thought it may be helpful to provide a summary of popular metrics here, for a better understanding of each metric and the applications they can be used for. In some applications looking at a single metric may not give you the whole picture of the problem you are solving, and you may want to use a subset of the metrics discussed in this post to have a concrete evaluation of your models. \n\n\n- [The MAP@5 metric](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304615) by @kishalmandal.\n\n\n<h1>Modeling</h1>\nThe most important step is to build a good model, that can well predict unseen data (aka Test Dataset), so choosing the right model is often not easy, it requires a lot of experiments and of course experience.\n\n- [Papers on Whale and Dolphin Identification 🐬🐋](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304973) by @datascientistfp.\n- [HappyWhale ArcFace Baseline (TPU)](https://www.kaggle.com/ks2019/happywhale-arcface-baseline-tpu) by @ks2019.\n- [[Pytorch] HappyWhale Siamese Starter](https://www.kaggle.com/debarshichanda/pytorch-happywhale-siamese-starter) by @debarshichanda.\n- [[Pytorch] ArcFace + GeM Pooling Starter](https://www.kaggle.com/debarshichanda/pytorch-arcface-gem-pooling-starter) by @debarshichanda.\n- [CNN with Keras Stater](https://www.kaggle.com/ammarnassanalhajali/cnn-with-keras-stater) by @ammarnassanalhajali.\n- [[Train][PyTorch] FineTune ConvNeXt](https://www.kaggle.com/ayuraj/train-pytorch-finetune-convnext) by @ayuraj.\n- [Species classification starter + TF data pipeline](https://www.kaggle.com/meowmeowmeowmeowmeow/species-classification-starter-tf-data-pipeline) by @meowmeowmeowmeowmeow.\n- [SOTA2020 CurricularFace+AMP+Mixup+CutMix+SAM](https://www.kaggle.com/sentrankim/sota2020-curricularface-amp-mixup-cutmix-sam) by @sentrankim.\n\n<h1>Resting</h1>\nTaking some rest for no long time is very important so you keep your brain fresh and avoid burnout. Also, it is useful to spend this period of time with the benefit of learning some additional materials for certain competitions.\n\n- [What do the species look like?](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305919) by @andradaolteanu.\n- [The most famous Whales!](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304509) by **me**.\n\n<h1>Intersting insights</h1>\n- [Duplicate names in species, can be merged together](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304633) by @karthickp6.\n- [LB probing and train/test split](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305428) by @olegsidorshin.\n- [Fin detect, extract and identify pipeline with paper and code](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305607) by @bsridatta.\n- [Avoid flipping the images!](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305843) by @bsridatta.\n- [Additional Metadata For Train/Test [Path and Image Size]](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305030) by @dschettler8845.\n- [Some images have more than one whale/Dolphin](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305899) by @chihantsai.\n- [Was Train Dataset applied Augmentations?](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304536) by **me**.",
      "votes": null
    },
    {
      "id": "1681895",
      "postDate": "02/08/2022 19:18:24",
      "content": "<p>Thanks for the amazing compilation <a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> !</p>",
      "rawMarkdown": "Thanks for the amazing compilation @vad13irt !",
      "votes": null
    },
    {
      "id": "1681903",
      "postDate": "02/08/2022 19:24:26",
      "content": "<p>You are welcome, <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>! You did great posts!</p>",
      "rawMarkdown": "You are welcome, @bsridatta! You did great posts!",
      "votes": null
    },
    {
      "id": "1681909",
      "postDate": "02/08/2022 19:30:24",
      "content": "<p>Thanks for mentioning them here. I shall try to add more useful observations!</p>",
      "rawMarkdown": "Thanks for mentioning them here. I shall try to add more useful observations!",
      "votes": null
    },
    {
      "id": "1684970",
      "postDate": "02/10/2022 22:34:38",
      "content": "<p><a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> Thanks for merging the links. Very useful, upvoted!</p>",
      "rawMarkdown": "vad13irt Thanks for merging the links. Very useful, upvoted!",
      "votes": null
    },
    {
      "id": "1685412",
      "postDate": "02/11/2022 08:51:55",
      "content": "<p>very interesting!</p>",
      "rawMarkdown": "very interesting!",
      "votes": null
    },
    {
      "id": "1685654",
      "postDate": "02/11/2022 12:55:55",
      "content": "<p>Thanks for the amazing compilation <a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> !!!!</p>",
      "rawMarkdown": "Thanks for the amazing compilation @vad13irt !!!!",
      "votes": null
    },
    {
      "id": "1685816",
      "postDate": "02/11/2022 15:01:53",
      "content": "<p>Thanks for Sharing <a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> </p>\n<p>I will Bookmark it 😎</p>",
      "rawMarkdown": "Thanks for Sharing @vad13irt \n\nI will Bookmark it 😎",
      "votes": null
    },
    {
      "id": "1686223",
      "postDate": "02/11/2022 21:26:13",
      "content": "<p>Impressive though I feel OpenCV or skimage  could've added value in EDA</p>",
      "rawMarkdown": "Impressive though I feel OpenCV or skimage  could've added value in EDA",
      "votes": null
    },
    {
      "id": "1689749",
      "postDate": "02/14/2022 13:24:31",
      "content": "<p>THANKS FOR GREEEEAT WORK!</p>",
      "rawMarkdown": "THANKS FOR GREEEEAT WORK!",
      "votes": null
    },
    {
      "id": "1700179",
      "postDate": "02/21/2022 18:12:03",
      "content": "<p>Great work, thanks a lot!</p>",
      "rawMarkdown": "Great work, thanks a lot!",
      "votes": null
    },
    {
      "id": "1700185",
      "postDate": "02/21/2022 18:18:14",
      "content": "<p>Congratulation on becoming GM :)</p>",
      "rawMarkdown": "Congratulation on becoming GM :)",
      "votes": null
    },
    {
      "id": "1700909",
      "postDate": "02/22/2022 11:21:41",
      "content": "<p>Great Work!!<br>\nThanks a lot</p>",
      "rawMarkdown": "Great Work!!\nThanks a lot",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1681895,
      "author_name": "bsridatta",
      "author_url": "",
      "post_date": "02/08/2022 19:18:24",
      "content": "<p>Thanks for the amazing compilation <a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> !</p>",
      "votes": null,
      "replies": [
        {
          "id": 1681903,
          "author_name": "vad13irt",
          "author_url": "",
          "post_date": "02/08/2022 19:24:26",
          "content": "<p>You are welcome, <a href=\"https://www.kaggle.com/bsridatta\" target=\"_blank\">@bsridatta</a>! You did great posts!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1681909,
          "author_name": "bsridatta",
          "author_url": "",
          "post_date": "02/08/2022 19:30:24",
          "content": "<p>Thanks for mentioning them here. I shall try to add more useful observations!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1684970,
      "author_name": "nebipeker",
      "author_url": "",
      "post_date": "02/10/2022 22:34:38",
      "content": "<p><a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> Thanks for merging the links. Very useful, upvoted!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1685412,
      "author_name": "akulardhala",
      "author_url": "",
      "post_date": "02/11/2022 08:51:55",
      "content": "<p>very interesting!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1685654,
      "author_name": "abdulmanankhalid",
      "author_url": "",
      "post_date": "02/11/2022 12:55:55",
      "content": "<p>Thanks for the amazing compilation <a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> !!!!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1685816,
      "author_name": "balasubramaniamv",
      "author_url": "",
      "post_date": "02/11/2022 15:01:53",
      "content": "<p>Thanks for Sharing <a href=\"https://www.kaggle.com/vad13irt\" target=\"_blank\">@vad13irt</a> </p>\n<p>I will Bookmark it 😎</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1686223,
      "author_name": "muhammadammarjamshed",
      "author_url": "",
      "post_date": "02/11/2022 21:26:13",
      "content": "<p>Impressive though I feel OpenCV or skimage  could've added value in EDA</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1689749,
      "author_name": "wuhaowang",
      "author_url": "",
      "post_date": "02/14/2022 13:24:31",
      "content": "<p>THANKS FOR GREEEEAT WORK!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1700179,
      "author_name": "datascientistfp",
      "author_url": "",
      "post_date": "02/21/2022 18:12:03",
      "content": "<p>Great work, thanks a lot!</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1700185,
      "author_name": "remekkinas",
      "author_url": "",
      "post_date": "02/21/2022 18:18:14",
      "content": "<p>Congratulation on becoming GM :)</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1700909,
      "author_name": "haruki741",
      "author_url": "",
      "post_date": "02/22/2022 11:21:41",
      "content": "<p>Great Work!!<br>\nThanks a lot</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1681821": "<h1>Exploratory Data Analysis</h1>\nAt beginning of the competition, you should do Exploratory Data Analysis, to find some relations in features and targets, write down some ideas for Feature Engineering from these explorations/insights, and much other stuff.\n\n\n- [Happywhale: Data Distribution 🐋🐬](https://www.kaggle.com/awsaf49/happywhale-data-distribution/) by @awsaf49.\n- [Happy Whales and Dolphins](https://www.kaggle.com/gpreda/happy-whales-and-dolphins#Images-data-exploration) by @gpreda.\n- [🐋 and 🐬 Identification: EDA + Augmentation](https://www.kaggle.com/ruchi798/and-identification-eda-augmentation) by @ruchi798.\n- [🐋🐬 STARTER+EDA+AUG](https://www.kaggle.com/abhranta/starter-eda-aug) by @abhranta.\n- [HappyWhale 🐳](https://www.kaggle.com/bsridatta/happywhale) by @bsridatta.\n- [Happywhale: Exploratory Data Analysis](https://www.kaggle.com/vad13irt/happywhale-exploratory-data-analysis/notebook) by **me**.\n- [Whale and Dolphin Identification EDA](https://www.kaggle.com/hamditarek/whale-and-dolphin-identification-eda) by @hamditarek.\n- [HappyWhale EDA,Fastai & Albumentations Starter🦈🐳](https://www.kaggle.com/ritesh2000/happywhale-eda-fastai-albumentations-starter) by @ritesh2000.\n- [😊🐳&🐬 - EDA and Baseline Solution](https://www.kaggle.com/dschettler8845/eda-and-baseline-solution) by @dschettler8845.\n\n\n<h1>Metric/Loss Understanding</h1>\nThe intermediate step, which also is one of the most important, is Metric Understanding. Choosing the right metric is crucial while evaluating machine learning (ML) models. Various metrics are proposed to evaluate ML models in different applications, and I thought it may be helpful to provide a summary of popular metrics here, for a better understanding of each metric and the applications they can be used for. In some applications looking at a single metric may not give you the whole picture of the problem you are solving, and you may want to use a subset of the metrics discussed in this post to have a concrete evaluation of your models. \n\n\n- [The MAP@5 metric](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304615) by @kishalmandal.\n\n\n<h1>Modeling</h1>\nThe most important step is to build a good model, that can well predict unseen data (aka Test Dataset), so choosing the right model is often not easy, it requires a lot of experiments and of course experience.\n\n- [Papers on Whale and Dolphin Identification 🐬🐋](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304973) by @datascientistfp.\n- [HappyWhale ArcFace Baseline (TPU)](https://www.kaggle.com/ks2019/happywhale-arcface-baseline-tpu) by @ks2019.\n- [[Pytorch] HappyWhale Siamese Starter](https://www.kaggle.com/debarshichanda/pytorch-happywhale-siamese-starter) by @debarshichanda.\n- [[Pytorch] ArcFace + GeM Pooling Starter](https://www.kaggle.com/debarshichanda/pytorch-arcface-gem-pooling-starter) by @debarshichanda.\n- [CNN with Keras Stater](https://www.kaggle.com/ammarnassanalhajali/cnn-with-keras-stater) by @ammarnassanalhajali.\n- [[Train][PyTorch] FineTune ConvNeXt](https://www.kaggle.com/ayuraj/train-pytorch-finetune-convnext) by @ayuraj.\n- [Species classification starter + TF data pipeline](https://www.kaggle.com/meowmeowmeowmeowmeow/species-classification-starter-tf-data-pipeline) by @meowmeowmeowmeowmeow.\n- [SOTA2020 CurricularFace+AMP+Mixup+CutMix+SAM](https://www.kaggle.com/sentrankim/sota2020-curricularface-amp-mixup-cutmix-sam) by @sentrankim.\n\n<h1>Resting</h1>\nTaking some rest for no long time is very important so you keep your brain fresh and avoid burnout. Also, it is useful to spend this period of time with the benefit of learning some additional materials for certain competitions.\n\n- [What do the species look like?](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305919) by @andradaolteanu.\n- [The most famous Whales!](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304509) by **me**.\n\n<h1>Intersting insights</h1>\n- [Duplicate names in species, can be merged together](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304633) by @karthickp6.\n- [LB probing and train/test split](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305428) by @olegsidorshin.\n- [Fin detect, extract and identify pipeline with paper and code](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305607) by @bsridatta.\n- [Avoid flipping the images!](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305843) by @bsridatta.\n- [Additional Metadata For Train/Test [Path and Image Size]](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305030) by @dschettler8845.\n- [Some images have more than one whale/Dolphin](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/305899) by @chihantsai.\n- [Was Train Dataset applied Augmentations?](https://www.kaggle.com/c/happy-whale-and-dolphin/discussion/304536) by **me**.",
    "1681895": "Thanks for the amazing compilation @vad13irt !",
    "1681903": "You are welcome, @bsridatta! You did great posts!",
    "1681909": "Thanks for mentioning them here. I shall try to add more useful observations!",
    "1684970": "vad13irt Thanks for merging the links. Very useful, upvoted!",
    "1685412": "very interesting!",
    "1685654": "Thanks for the amazing compilation @vad13irt !!!!",
    "1685816": "Thanks for Sharing @vad13irt \n\nI will Bookmark it 😎",
    "1686223": "Impressive though I feel OpenCV or skimage  could've added value in EDA",
    "1689749": "THANKS FOR GREEEEAT WORK!",
    "1700179": "Great work, thanks a lot!",
    "1700185": "Congratulation on becoming GM :)",
    "1700909": "Great Work!!\nThanks a lot"
  },
  "source": "meta"
}