{
  "id": 166295,
  "title": "How to stabilize LB score for the same Experiment?",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/166295",
  "author_name": "Durgesh",
  "post_date": "2020-07-12T12:18:39.033000",
  "votes": 0,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Despite of seeding everything properly[I guess]. \nLB Score is not recreating itself while running notebook with same configuration.</p>\n\n<p><strong>Notebook:</strong> <br>\n* <a href=\"https://www.kaggle.com/orionpax00/melanoma-detection-single-model-gpu-tpu-yapl?scriptVersionId=38533975\">NOTEBOOK 1</a> : LB 0.874\n* <a href=\"https://www.kaggle.com/orionpax00/fork-of-melanoma-detection-single-model-gpu-79ddb7\">NOTEBOOK 2</a> : LB 0.862</p>\n\n<p>Any Suggestion about how to Stabilize LB Score.</p>\n\n<h3>Experiment Configuration</h3>\n\n<ul>\n<li>128x128 size Images <a href=\"https://www.kaggle.com/cdeotte/melanoma-128x128\">LINK</a> </li>\n<li>5 fold Training</li>\n<li>EfficientnetB0</li>\n<li>Seed : 42</li>\n<li>Augmentation\n<ul><li>random flip left right</li>\n<li>rot90</li></ul></li>\n<li>LR Scheduler : cosine schedule with warmup</li>\n</ul>",
  "messages": [
    {
      "id": 925988,
      "postDate": "2020-07-12T12:18:39.033Z",
      "content": "<p>Despite of seeding everything properly[I guess]. \nLB Score is not recreating itself while running notebook with same configuration.</p>\n\n<p><strong>Notebook:</strong> <br>\n* <a href=\"https://www.kaggle.com/orionpax00/melanoma-detection-single-model-gpu-tpu-yapl?scriptVersionId=38533975\">NOTEBOOK 1</a> : LB 0.874\n* <a href=\"https://www.kaggle.com/orionpax00/fork-of-melanoma-detection-single-model-gpu-79ddb7\">NOTEBOOK 2</a> : LB 0.862</p>\n\n<p>Any Suggestion about how to Stabilize LB Score.</p>\n\n<h3>Experiment Configuration</h3>\n\n<ul>\n<li>128x128 size Images <a href=\"https://www.kaggle.com/cdeotte/melanoma-128x128\">LINK</a> </li>\n<li>5 fold Training</li>\n<li>EfficientnetB0</li>\n<li>Seed : 42</li>\n<li>Augmentation\n<ul><li>random flip left right</li>\n<li>rot90</li></ul></li>\n<li>LR Scheduler : cosine schedule with warmup</li>\n</ul>",
      "rawMarkdown": "Despite of seeding everything properly[I guess]. \nLB Score is not recreating itself while running notebook with same configuration.\n\n**Notebook:** <br>\n* [NOTEBOOK 1](https://www.kaggle.com/orionpax00/melanoma-detection-single-model-gpu-tpu-yapl?scriptVersionId=38533975) : LB 0.874\n* [NOTEBOOK 2](https://www.kaggle.com/orionpax00/fork-of-melanoma-detection-single-model-gpu-79ddb7) : LB 0.862\n\nAny Suggestion about how to Stabilize LB Score.\n\n\n### Experiment Configuration\n* 128x128 size Images [LINK](https://www.kaggle.com/cdeotte/melanoma-128x128) \n* 5 fold Training\n* EfficientnetB0\n* Seed : 42\n* Augmentation\n   * random flip left right\n   * rot90\n* LR Scheduler : cosine schedule with warmup"
    },
    {
      "id": 926711,
      "postDate": "2020-07-12T23:19:51.587Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 926711,
      "author_name": "",
      "author_url": "",
      "post_date": "2020-07-12T23:19:51.587000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "925988": "Despite of seeding everything properly[I guess]. \nLB Score is not recreating itself while running notebook with same configuration.\n\n**Notebook:** <br>\n* [NOTEBOOK 1](https://www.kaggle.com/orionpax00/melanoma-detection-single-model-gpu-tpu-yapl?scriptVersionId=38533975) : LB 0.874\n* [NOTEBOOK 2](https://www.kaggle.com/orionpax00/fork-of-melanoma-detection-single-model-gpu-79ddb7) : LB 0.862\n\nAny Suggestion about how to Stabilize LB Score.\n\n\n### Experiment Configuration\n* 128x128 size Images [LINK](https://www.kaggle.com/cdeotte/melanoma-128x128) \n* 5 fold Training\n* EfficientnetB0\n* Seed : 42\n* Augmentation\n   * random flip left right\n   * rot90\n* LR Scheduler : cosine schedule with warmup",
    "926711": ""
  }
}