{
  "id": 170075,
  "title": "Uncontrolled RAM growth - SIIM ISIC Melanoma",
  "url": "/competitions/siim-isic-melanoma-classification/discussion/170075",
  "author_name": "Aditya Baurai",
  "post_date": "2020-07-26T11:04:13.277000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2599569%2F2f9ddcfaa777ef07526661a50e9ae44d%2FScreenshot%20(152\" alt=\"\">.png?generation=1595761320750145&amp;alt=media)</p>\n\n<p>I have been using MobileNetV2(and previously tried with ResNet). Just observe after one epoch the RAM growth is extremely large and it goes out of memory.</p>\n\n<ul>\n<li>Batch size = 8 </li>\n<li>Used tf.data API</li>\n</ul>\n\n<p>What's going wrong and where ? </p>\n\n<p>Even in check points I have diabled save model weights ModelCheckPoint callback. What am I missing? Can anyone aid in this crisis??</p>",
  "messages": [
    {
      "id": 946094,
      "postDate": "2020-07-26T11:04:13.277Z",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2599569%2F2f9ddcfaa777ef07526661a50e9ae44d%2FScreenshot%20(152\" alt=\"\">.png?generation=1595761320750145&amp;alt=media)</p>\n\n<p>I have been using MobileNetV2(and previously tried with ResNet). Just observe after one epoch the RAM growth is extremely large and it goes out of memory.</p>\n\n<ul>\n<li>Batch size = 8 </li>\n<li>Used tf.data API</li>\n</ul>\n\n<p>What's going wrong and where ? </p>\n\n<p>Even in check points I have diabled save model weights ModelCheckPoint callback. What am I missing? Can anyone aid in this crisis??</p>",
      "rawMarkdown": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2599569%2F2f9ddcfaa777ef07526661a50e9ae44d%2FScreenshot%20(152).png?generation=1595761320750145&amp;alt=media)\n\nI have been using MobileNetV2(and previously tried with ResNet). Just observe after one epoch the RAM growth is extremely large and it goes out of memory.\n\n- Batch size = 8 \n- Used tf.data API\n\nWhat's going wrong and where ? \n\nEven in check points I have diabled save model weights ModelCheckPoint callback. What am I missing? Can anyone aid in this crisis??"
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "946094": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F2599569%2F2f9ddcfaa777ef07526661a50e9ae44d%2FScreenshot%20(152).png?generation=1595761320750145&amp;alt=media)\n\nI have been using MobileNetV2(and previously tried with ResNet). Just observe after one epoch the RAM growth is extremely large and it goes out of memory.\n\n- Batch size = 8 \n- Used tf.data API\n\nWhat's going wrong and where ? \n\nEven in check points I have diabled save model weights ModelCheckPoint callback. What am I missing? Can anyone aid in this crisis??"
  }
}