{
  "id": 301239,
  "title": "yolov5 training, the output of training info not complete problem",
  "url": "/competitions/tensorflow-great-barrier-reef/discussion/301239",
  "author_name": "",
  "post_date": "2022-01-16T15:03:18.026362900Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>my forked notebook:   missing P R  mAP etc info</p>\n<p>train: Scanning '../yolo_data/fold1/labels/train' images and labels…2820 found<br>\nval: Scanning '../yolo_data/fold1/labels/val' images and labels…2099 found, 61<br>\n       0/7     13.6G   0.07514   0.08973         0         1      3648: 100%|███<br>\n               Class     Images     Labels          P          R     mAP@.5 mAP@<br>\n       1/7     13.6G   0.05326    0.0675         0        11      3648: 100%|███<br>\n               Class     Images     Labels          P          R     mAP@.5 mAP@</p>\n<hr>\n<p>while the original notebook output:</p>\n<pre><code> Epoch   gpu_mem       box       obj       cls    labels  img_size\n   0/4     7.11G    0.0777   0.08275         0        14      3008: 100%|███\n           Class     Images     Labels          P          R     mAP@.5 mAP@\n             all       8232       6384     0.0103     0.0426    0.00205   0.000597\n\n Epoch   gpu_mem       box       obj       cls    labels  img_size\n   1/4       16G   0.05507   0.06298         0        17      3008: 100%|███\n           Class     Images     Labels          P          R     mAP@.5 mAP@\n             all       8232       6384      0.386      0.364      0.299     0.0661\n</code></pre>\n<p>btw, on colab,  the P R mAP etc do show up, however not on Kaggle for my forked notebook.</p>",
  "messages": [
    {
      "id": "1652384",
      "postDate": "01/16/2022 15:03:18",
      "content": "<p>my forked notebook:   missing P R  mAP etc info</p>\n<p>train: Scanning '../yolo_data/fold1/labels/train' images and labels…2820 found<br>\nval: Scanning '../yolo_data/fold1/labels/val' images and labels…2099 found, 61<br>\n       0/7     13.6G   0.07514   0.08973         0         1      3648: 100%|███<br>\n               Class     Images     Labels          P          R     mAP@.5 mAP@<br>\n       1/7     13.6G   0.05326    0.0675         0        11      3648: 100%|███<br>\n               Class     Images     Labels          P          R     mAP@.5 mAP@</p>\n<hr>\n<p>while the original notebook output:</p>\n<pre><code> Epoch   gpu_mem       box       obj       cls    labels  img_size\n   0/4     7.11G    0.0777   0.08275         0        14      3008: 100%|███\n           Class     Images     Labels          P          R     mAP@.5 mAP@\n             all       8232       6384     0.0103     0.0426    0.00205   0.000597\n\n Epoch   gpu_mem       box       obj       cls    labels  img_size\n   1/4       16G   0.05507   0.06298         0        17      3008: 100%|███\n           Class     Images     Labels          P          R     mAP@.5 mAP@\n             all       8232       6384      0.386      0.364      0.299     0.0661\n</code></pre>\n<p>btw, on colab,  the P R mAP etc do show up, however not on Kaggle for my forked notebook.</p>",
      "rawMarkdown": "my forked notebook:   missing P R  mAP etc info\n\ntrain: Scanning '../yolo_data/fold1/labels/train' images and labels...2820 found\nval: Scanning '../yolo_data/fold1/labels/val' images and labels...2099 found, 61\n       0/7     13.6G   0.07514   0.08973         0         1      3648: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n       1/7     13.6G   0.05326    0.0675         0        11      3648: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n  \n******************************************************************\nwhile the original notebook output:\n\n     Epoch   gpu_mem       box       obj       cls    labels  img_size\n       0/4     7.11G    0.0777   0.08275         0        14      3008: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n                 all       8232       6384     0.0103     0.0426    0.00205   0.000597\n\n     Epoch   gpu_mem       box       obj       cls    labels  img_size\n       1/4       16G   0.05507   0.06298         0        17      3008: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n                 all       8232       6384      0.386      0.364      0.299     0.0661\n\n\nbtw, on colab,  the P R mAP etc do show up, however not on Kaggle for my forked notebook.",
      "votes": null
    },
    {
      "id": "1652451",
      "postDate": "01/16/2022 16:25:14",
      "content": "<p>I faced the same problem.<br>\nHowever, you can check the result from the output file \"result.csv\".</p>",
      "rawMarkdown": "I faced the same problem.\nHowever, you can check the result from the output file \"result.csv\".",
      "votes": null
    },
    {
      "id": "1652490",
      "postDate": "01/16/2022 17:15:28",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/dragonzhang\" target=\"_blank\">@dragonzhang</a> you should fork the original repo yolov5 and add line code below line 268 in val.py (yolov5 repo):<br>\nprint(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))<br>\nYou will see the result after an epoch. Kaggle notebook sometimes gets this issue. Hope its help</p>",
      "rawMarkdown": "Hi @dragonzhang you should fork the original repo yolov5 and add line code below line 268 in val.py (yolov5 repo):\nprint(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))\nYou will see the result after an epoch. Kaggle notebook sometimes gets this issue. Hope its help",
      "votes": null
    },
    {
      "id": "1652865",
      "postDate": "01/17/2022 04:18:15",
      "content": "<p>thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.</p>",
      "rawMarkdown": "thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.",
      "votes": null
    },
    {
      "id": "1652866",
      "postDate": "01/17/2022 04:18:26",
      "content": "<p>thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.</p>",
      "rawMarkdown": "thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.",
      "votes": null
    },
    {
      "id": "1652869",
      "postDate": "01/17/2022 04:19:33",
      "content": "<p>thank you for your reply.</p>",
      "rawMarkdown": "thank you for your reply.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1652451,
      "author_name": "syurenuko",
      "author_url": "",
      "post_date": "01/16/2022 16:25:14",
      "content": "<p>I faced the same problem.<br>\nHowever, you can check the result from the output file \"result.csv\".</p>",
      "votes": null,
      "replies": [
        {
          "id": 1652869,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "01/17/2022 04:19:33",
          "content": "<p>thank you for your reply.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1652490,
      "author_name": "magiccard",
      "author_url": "",
      "post_date": "01/16/2022 17:15:28",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/dragonzhang\" target=\"_blank\">@dragonzhang</a> you should fork the original repo yolov5 and add line code below line 268 in val.py (yolov5 repo):<br>\nprint(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))<br>\nYou will see the result after an epoch. Kaggle notebook sometimes gets this issue. Hope its help</p>",
      "votes": null,
      "replies": [
        {
          "id": 1652865,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "01/17/2022 04:18:15",
          "content": "<p>thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1652866,
          "author_name": "dragonzhang",
          "author_url": "",
          "post_date": "01/17/2022 04:18:26",
          "content": "<p>thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1652384": "my forked notebook:   missing P R  mAP etc info\n\ntrain: Scanning '../yolo_data/fold1/labels/train' images and labels...2820 found\nval: Scanning '../yolo_data/fold1/labels/val' images and labels...2099 found, 61\n       0/7     13.6G   0.07514   0.08973         0         1      3648: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n       1/7     13.6G   0.05326    0.0675         0        11      3648: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n  \n******************************************************************\nwhile the original notebook output:\n\n     Epoch   gpu_mem       box       obj       cls    labels  img_size\n       0/4     7.11G    0.0777   0.08275         0        14      3008: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n                 all       8232       6384     0.0103     0.0426    0.00205   0.000597\n\n     Epoch   gpu_mem       box       obj       cls    labels  img_size\n       1/4       16G   0.05507   0.06298         0        17      3008: 100%|███\n               Class     Images     Labels          P          R     mAP@.5 mAP@\n                 all       8232       6384      0.386      0.364      0.299     0.0661\n\n\nbtw, on colab,  the P R mAP etc do show up, however not on Kaggle for my forked notebook.",
    "1652451": "I faced the same problem.\nHowever, you can check the result from the output file \"result.csv\".",
    "1652490": "Hi @dragonzhang you should fork the original repo yolov5 and add line code below line 268 in val.py (yolov5 repo):\nprint(pf % (names[c], seen, nt[c], p[i], r[i], ap50[i], ap[i]))\nYou will see the result after an epoch. Kaggle notebook sometimes gets this issue. Hope its help",
    "1652865": "thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.",
    "1652866": "thank you for your reply.   I run same notebook on colab pro,  only dataset directories changed,  no this problem. that why I ask.",
    "1652869": "thank you for your reply."
  },
  "source": "meta"
}