[Bug] Images disappear with a large number of labels in Image Classification

[Foldblade]

English is not my native language; please excuse typing errors.

The bug: When using Image Classification, images may disappear with a large number of labels. In the following image, you can see the right scrollbars are on the top, but the image is disappear.

Steps to Reproduce? Using this following JSON, and then preview or labeling it.

An example JSON Dataset

{
  "name": "New undefined Dataset",
  "interface": {
    "type": "image_classification",
    "labels": [
      "瓷器",
      "白瓷",
      "青瓷",
      "青花",
      "五彩",
      "斗彩",
      "粉彩",
      "广彩",
      "克拉克瓷",
      "唐三彩",
      "釉里红",
      "邢窑",
      "越窑",
      "汝窑",
      "官窑（宋）",
      "哥窑",
      "钧窑",
      "定窑",
      "龙泉窑",
      "长沙窑",
      "漳州窑",
      "耀州窑",
      "磁州窑",
      "德化窑",
      "吉州窑",
      "建盏",
      "陶器",
      "彩陶",
      "灰陶",
      "红陶",
      "黑陶",
      "印纹硬陶",
      "原始瓷器",
      "古籍_善本",
      "书法_绘画_书画",
      "家具",
      "宝石",
      "文具",
      "标本_化石",
      "武器",
      "法器",
      "漆器",
      "牙骨角器",
      "玉石器",
      "玺_印_符_牌",
      "玻璃器",
      "珐琅器",
      "甲骨",
      "石器_石刻_砖瓦",
      "碑_帖_拓本",
      "竹雕_木雕",
      "织绣",
      "邮品",
      "金银器",
      "钱币",
      "铁器",
      "铜器",
      "青铜器",
      "雕塑_造像",
      "匜",
      "卣",
      "尊",
      "敦",
      "斝",
      "爵",
      "瓿",
      "甑",
      "甗",
      "盉",
      "盨",
      "簋",
      "簠",
      "罍",
      "角",
      "觚",
      "觥",
      "觯",
      "豆",
      "釜",
      "鬲",
      "鬶",
      "鼎",
      "蒜头瓶",
      "天球瓶",
      "纸槌瓶",
      "高足杯碗"
    ]
  },
  "samples": [
    {
      "_id": "s2k0qa7do",
      "imageUrl": "https://upload.wikimedia.org/wikipedia/commons/thumb/0/01/Alcedo_atthis_-_Riserve_naturali_e_aree_contigue_della_fascia_fluviale_del_Po.jpg/1280px-Alcedo_atthis_-_Riserve_naturali_e_aree_contigue_della_fascia_fluviale_del_Po.jpg"
    }
  ]
}

---

I tried to fix this problem in #534.

Fixing this helps a lot when labeling my Image Classification dataset which has a large number of labels. May you check the pull request, please? Thank you!