{
  "id": 163290,
  "title": "Dataloader (WIP)",
  "url": "/competitions/landmark-retrieval-2020/discussion/163290",
  "author_name": "Trigram",
  "post_date": "2020-07-01T15:19:20.180000",
  "votes": 3,
  "comment_count": 0,
  "views": 0,
  "content": "<p><code>python\npaths=[]\nY=[]\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n</code></p>\n\n<p>Now if you want a convert to tensor method:\n<code>python\npaths=[]\nY=[]\ntensors = []\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                image = plt.imread(y4)\n                tensor = tf.convert_to_tensor(image)\n                tensors.append(tensor)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n</code></p>\n\n<p>It is fairly runnable on Kaggle notebooks.</p>",
  "messages": [
    {
      "id": 911118,
      "postDate": "2020-07-01T15:19:20.180Z",
      "content": "<p><code>python\npaths=[]\nY=[]\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n</code></p>\n\n<p>Now if you want a convert to tensor method:\n<code>python\npaths=[]\nY=[]\ntensors = []\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                image = plt.imread(y4)\n                tensor = tf.convert_to_tensor(image)\n                tensors.append(tensor)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n</code></p>\n\n<p>It is fairly runnable on Kaggle notebooks.</p>",
      "rawMarkdown": "```python\npaths=[]\nY=[]\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n```\n\nNow if you want a convert to tensor method:\n```python\npaths=[]\nY=[]\ntensors = []\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                image = plt.imread(y4)\n                tensor = tf.convert_to_tensor(image)\n                tensors.append(tensor)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n```\n\nIt is fairly runnable on Kaggle notebooks.",
      "votes": 3
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "911118": "```python\npaths=[]\nY=[]\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n```\n\nNow if you want a convert to tensor method:\n```python\npaths=[]\nY=[]\ntensors = []\nfor y1 in os.listdir('../input/landmark-retrieval-2020/train/'):\n    for y2 in os.listdir('../input/landmark-retrieval-2020/train/' + y1):\n        for y3 in os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2):\n            for y4 in tqdm(os.listdir('../input/landmark-retrieval-2020/train/' + y1 + '/' + y2  + '/' + y3)):\n                paths.append(y4)\n                image = plt.imread(y4)\n                tensor = tf.convert_to_tensor(image)\n                tensors.append(tensor)\n                if len(paths) == 1580470:\n                    raise KeyboardInterrupt\n                    break\n```\n\nIt is fairly runnable on Kaggle notebooks."
  }
}