{
  "id": 264068,
  "title": "FastAI: No stats displayed when using One Cycle",
  "url": "/competitions/rsna-miccai-brain-tumor-radiogenomic-classification/discussion/264068",
  "author_name": "",
  "post_date": "2021-08-11T02:19:00.359781500Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<pre><code>learn.fit_one_cycle(10, lr_max=1e-2, cbs=ShortEpochCallback())\n\nepoch    train_loss  valid_loss  error_rate  accuracy    time\n0    00:01\n1    00:01\n2    00:01\n3    00:01\n4    00:01\n5    00:01\n6    00:01\n7    00:01\n8    00:01\n9    00:01\n</code></pre>\n<p>When running <code>fit_one_cycle</code>, I don't get any valid_loss, error_rate, nor accuracy_time. Does anyone have experience with this problem? In examples online I can see these columns are filled in. </p>",
  "messages": [
    {
      "id": "1465319",
      "postDate": "08/11/2021 02:19:00",
      "content": "<pre><code>learn.fit_one_cycle(10, lr_max=1e-2, cbs=ShortEpochCallback())\n\nepoch    train_loss  valid_loss  error_rate  accuracy    time\n0    00:01\n1    00:01\n2    00:01\n3    00:01\n4    00:01\n5    00:01\n6    00:01\n7    00:01\n8    00:01\n9    00:01\n</code></pre>\n<p>When running <code>fit_one_cycle</code>, I don't get any valid_loss, error_rate, nor accuracy_time. Does anyone have experience with this problem? In examples online I can see these columns are filled in. </p>",
      "rawMarkdown": "```\nlearn.fit_one_cycle(10, lr_max=1e-2, cbs=ShortEpochCallback())\n\nepoch\ttrain_loss\tvalid_loss\terror_rate\taccuracy\ttime\n0\t00:01\n1\t00:01\n2\t00:01\n3\t00:01\n4\t00:01\n5\t00:01\n6\t00:01\n7\t00:01\n8\t00:01\n9\t00:01\n```\n\nWhen running `fit_one_cycle`, I don't get any valid_loss, error_rate, nor accuracy_time. Does anyone have experience with this problem? In examples online I can see these columns are filled in.",
      "votes": null
    },
    {
      "id": "1466461",
      "postDate": "08/11/2021 13:28:27",
      "content": "<p>Maybe need more context</p>",
      "rawMarkdown": "Maybe need more context",
      "votes": null
    },
    {
      "id": "1467374",
      "postDate": "08/12/2021 00:31:34",
      "content": "<p><a href=\"https://github.com/fastai/fastai/issues/3126\" target=\"_blank\">https://github.com/fastai/fastai/issues/3126</a></p>\n<p>For context, this is basically the same problem I am having. It turns out passing a <code>short_valid=False</code> parameter presents some results.</p>",
      "rawMarkdown": "https://github.com/fastai/fastai/issues/3126\n\nFor context, this is basically the same problem I am having. It turns out passing a `short_valid=False` parameter presents some results.",
      "votes": null
    },
    {
      "id": "1467381",
      "postDate": "08/12/2021 00:38:42",
      "content": "<p>I think the best course of action is just remove the use of the Short Epoch Callback.</p>",
      "rawMarkdown": "I think the best course of action is just remove the use of the Short Epoch Callback.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1466461,
      "author_name": "marshath",
      "author_url": "",
      "post_date": "08/11/2021 13:28:27",
      "content": "<p>Maybe need more context</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1467374,
      "author_name": "d223chen",
      "author_url": "",
      "post_date": "08/12/2021 00:31:34",
      "content": "<p><a href=\"https://github.com/fastai/fastai/issues/3126\" target=\"_blank\">https://github.com/fastai/fastai/issues/3126</a></p>\n<p>For context, this is basically the same problem I am having. It turns out passing a <code>short_valid=False</code> parameter presents some results.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1467381,
      "author_name": "d223chen",
      "author_url": "",
      "post_date": "08/12/2021 00:38:42",
      "content": "<p>I think the best course of action is just remove the use of the Short Epoch Callback.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1465319": "```\nlearn.fit_one_cycle(10, lr_max=1e-2, cbs=ShortEpochCallback())\n\nepoch\ttrain_loss\tvalid_loss\terror_rate\taccuracy\ttime\n0\t00:01\n1\t00:01\n2\t00:01\n3\t00:01\n4\t00:01\n5\t00:01\n6\t00:01\n7\t00:01\n8\t00:01\n9\t00:01\n```\n\nWhen running `fit_one_cycle`, I don't get any valid_loss, error_rate, nor accuracy_time. Does anyone have experience with this problem? In examples online I can see these columns are filled in.",
    "1466461": "Maybe need more context",
    "1467374": "https://github.com/fastai/fastai/issues/3126\n\nFor context, this is basically the same problem I am having. It turns out passing a `short_valid=False` parameter presents some results.",
    "1467381": "I think the best course of action is just remove the use of the Short Epoch Callback."
  },
  "source": "meta"
}