{
  "id": 123948,
  "title": "Anyone using Caffe?",
  "url": "/competitions/deepfake-detection-challenge/discussion/123948",
  "author_name": "Baby Pants",
  "post_date": "2019-12-31T19:56:18.677000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello, I'm new to Kaggle and Python so I apologize in advance if my questions are stupid.\nI'm trying to import a custom network I trained using dlib in C++. </p>\n\n<p>As far as I can tell the only way to do this is to (1) export my network to XML from dlib; (2) run the \"convert_dlib_to_caffe\" conversion tool that's provided with dlib; (3) use the created Python module to save my model as a Caffe model files; and finally (4) load and run these files as a Caffe network...</p>\n\n<p>This method requires that I be able to use Caffe from a Kaggle notebook. Is that possible and if so how do I install it? If that's not possible (I saw no reference to Caffe in these various discussions) is there a different way for me to run my network in a Kaggle notebook?</p>\n\n<p>I was able to install dlib after adding the \"deepfake-models-packages\" to my data and I'm hoping I can do something similar for Caffe.</p>\n\n<p>Thanks in advance.</p>",
  "messages": [
    {
      "id": 707426,
      "postDate": "2019-12-31T19:56:18.677Z",
      "content": "<p>Hello, I'm new to Kaggle and Python so I apologize in advance if my questions are stupid.\nI'm trying to import a custom network I trained using dlib in C++. </p>\n\n<p>As far as I can tell the only way to do this is to (1) export my network to XML from dlib; (2) run the \"convert_dlib_to_caffe\" conversion tool that's provided with dlib; (3) use the created Python module to save my model as a Caffe model files; and finally (4) load and run these files as a Caffe network...</p>\n\n<p>This method requires that I be able to use Caffe from a Kaggle notebook. Is that possible and if so how do I install it? If that's not possible (I saw no reference to Caffe in these various discussions) is there a different way for me to run my network in a Kaggle notebook?</p>\n\n<p>I was able to install dlib after adding the \"deepfake-models-packages\" to my data and I'm hoping I can do something similar for Caffe.</p>\n\n<p>Thanks in advance.</p>",
      "rawMarkdown": "Hello, I'm new to Kaggle and Python so I apologize in advance if my questions are stupid.\nI'm trying to import a custom network I trained using dlib in C++. \n\nAs far as I can tell the only way to do this is to (1) export my network to XML from dlib; (2) run the \"convert_dlib_to_caffe\" conversion tool that's provided with dlib; (3) use the created Python module to save my model as a Caffe model files; and finally (4) load and run these files as a Caffe network...\n\nThis method requires that I be able to use Caffe from a Kaggle notebook. Is that possible and if so how do I install it? If that's not possible (I saw no reference to Caffe in these various discussions) is there a different way for me to run my network in a Kaggle notebook?\n\nI was able to install dlib after adding the \"deepfake-models-packages\" to my data and I'm hoping I can do something similar for Caffe.\n\nThanks in advance."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "707426": "Hello, I'm new to Kaggle and Python so I apologize in advance if my questions are stupid.\nI'm trying to import a custom network I trained using dlib in C++. \n\nAs far as I can tell the only way to do this is to (1) export my network to XML from dlib; (2) run the \"convert_dlib_to_caffe\" conversion tool that's provided with dlib; (3) use the created Python module to save my model as a Caffe model files; and finally (4) load and run these files as a Caffe network...\n\nThis method requires that I be able to use Caffe from a Kaggle notebook. Is that possible and if so how do I install it? If that's not possible (I saw no reference to Caffe in these various discussions) is there a different way for me to run my network in a Kaggle notebook?\n\nI was able to install dlib after adding the \"deepfake-models-packages\" to my data and I'm hoping I can do something similar for Caffe.\n\nThanks in advance."
  }
}