This competition is a No internet competition. This means that when you are uploading a submission your notebook will not be able to access the internet.
\nThis may be an issue for you if you need to access the internet, for example, if you need to download a GitHub repo to use as a library.
\nHow can we overcome this?
\nEasy, download the repo as a zip file and upload it as a dataset.
\nFor example, I needed to use this EfficientNet PyTorch repo, so I could not download it at inference time. Just upload it as a dataset and run:
\nimport sys \n\nsys.path.append(\"../input/efficientnet-pytorch/EfficientNet-PyTorch-master\") \n\nimport efficientnet_pytorch\nVoilà, you have imported a GitHub repo without internet!
", "messages": [ { "id": 1855706, "postDate": "2022-07-14T20:14:37.407Z", "content": "This competition is a No internet competition. This means that when you are uploading a submission your notebook will not be able to access the internet.
\nThis may be an issue for you if you need to access the internet, for example, if you need to download a GitHub repo to use as a library.
\nHow can we overcome this?
\nEasy, download the repo as a zip file and upload it as a dataset.
\nFor example, I needed to use this EfficientNet PyTorch repo, so I could not download it at inference time. Just upload it as a dataset and run:
\nimport sys \n\nsys.path.append(\"../input/efficientnet-pytorch/EfficientNet-PyTorch-master\") \n\nimport efficientnet_pytorch\nVoilà, you have imported a GitHub repo without internet!
", "rawMarkdown": "This competition is a **No internet** competition. This means that when you are uploading a submission your notebook will not be able to access the internet.\n\nThis may be an issue for you if you need to access the internet, for example, if you need to **download a GitHub repo to use as a library.**\n\nHow can we overcome this?\n\nEasy, download the repo as a zip file and upload it as a dataset. \n\nFor example, I needed to use this [EfficientNet PyTorch repo](https://github.com/lukemelas/EfficientNet-PyTorch), so I could not download it at inference time. Just upload it as a dataset and run:\n\n```\nimport sys \n\nsys.path.append(\"../input/efficientnet-pytorch/EfficientNet-PyTorch-master\") \n\nimport efficientnet_pytorch\n```\n\nVoilà, you have imported a GitHub repo without internet!", "votes": 6 }, { "id": 1859203, "postDate": "2022-07-17T13:27:42.050Z", "content": "I made a notebook to show how to install packages offline. Check it out! https://www.kaggle.com/code/analokamus/how-to-use-pyvips-offline/notebook
", "rawMarkdown": "I made a notebook to show how to install packages offline. Check it out! https://www.kaggle.com/code/analokamus/how-to-use-pyvips-offline/notebook", "votes": 1 }, { "id": 1857479, "postDate": "2022-07-16T06:33:06.487Z", "content": "Good idea, this could be useful to many of us while submitting our work!
", "rawMarkdown": "Good idea, this could be useful to many of us while submitting our work!", "votes": 1, "replies": [ { "id": 1858113, "postDate": "2022-07-16T17:09:37.797Z", "content": "Indeed Ravi. At inference time we cannot access internet and if you rely on libraries and packages that do not come built in the notebook you need this approach.
", "rawMarkdown": "Indeed Ravi. At inference time we cannot access internet and if you rely on libraries and packages that do not come built in the notebook you need this approach." } ] }, { "id": 1868040, "postDate": "2022-07-23T16:36:30.193Z", "content": "I wrote a blog about it how you can use one notebook as a dataset for another kernel that is running offline:
\nhttps://towardsdatascience.com/easy-kaggle-offline-submission-with-chaining-kernels-30bba5ea5c4d
Thanks for the great help Jirka, love your work :)
", "rawMarkdown": "Thanks for the great help Jirka, love your work :)" } ] }, { "id": 1858308, "postDate": "2022-07-16T19:50:45.883Z", "content": "This is a great tip! I had the same issue with trying to load some pretrained weights. I ended up just downloading the files and compressing them as ZIP file, uploading it to my Kaggle account, and then importing it like you said. Worked like a charm!
", "rawMarkdown": "This is a great tip! I had the same issue with trying to load some pretrained weights. I ended up just downloading the files and compressing them as ZIP file, uploading it to my Kaggle account, and then importing it like you said. Worked like a charm!\n" }, { "id": 1872152, "postDate": "2022-07-26T17:45:10.340Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 1859203, "author_name": "RabotniKuma", "author_url": "", "post_date": "2022-07-17T13:27:42.050000", "content": "I made a notebook to show how to install packages offline. Check it out! https://www.kaggle.com/code/analokamus/how-to-use-pyvips-offline/notebook
", "votes": 1, "replies": [] }, { "id": 1857479, "author_name": "Ravi Ramakrishnan", "author_url": "", "post_date": "2022-07-16T06:33:06.487000", "content": "Good idea, this could be useful to many of us while submitting our work!
", "votes": 1, "replies": [ { "id": 1858113, "author_name": "moth", "author_url": "", "post_date": "2022-07-16T17:09:37.797000", "content": "Indeed Ravi. At inference time we cannot access internet and if you rely on libraries and packages that do not come built in the notebook you need this approach.
", "votes": 0, "replies": [] } ] }, { "id": 1868040, "author_name": "Jirka", "author_url": "", "post_date": "2022-07-23T16:36:30.193000", "content": "I wrote a blog about it how you can use one notebook as a dataset for another kernel that is running offline:
\nhttps://towardsdatascience.com/easy-kaggle-offline-submission-with-chaining-kernels-30bba5ea5c4d
Thanks for the great help Jirka, love your work :)
", "votes": 0, "replies": [] } ] }, { "id": 1858308, "author_name": "The Devastator", "author_url": "", "post_date": "2022-07-16T19:50:45.883000", "content": "This is a great tip! I had the same issue with trying to load some pretrained weights. I ended up just downloading the files and compressing them as ZIP file, uploading it to my Kaggle account, and then importing it like you said. Worked like a charm!
", "votes": 0, "replies": [] }, { "id": 1872152, "author_name": "", "author_url": "", "post_date": "2022-07-26T17:45:10.340000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "1855706": "This competition is a **No internet** competition. This means that when you are uploading a submission your notebook will not be able to access the internet.\n\nThis may be an issue for you if you need to access the internet, for example, if you need to **download a GitHub repo to use as a library.**\n\nHow can we overcome this?\n\nEasy, download the repo as a zip file and upload it as a dataset. \n\nFor example, I needed to use this [EfficientNet PyTorch repo](https://github.com/lukemelas/EfficientNet-PyTorch), so I could not download it at inference time. Just upload it as a dataset and run:\n\n```\nimport sys \n\nsys.path.append(\"../input/efficientnet-pytorch/EfficientNet-PyTorch-master\") \n\nimport efficientnet_pytorch\n```\n\nVoilà, you have imported a GitHub repo without internet!", "1859203": "I made a notebook to show how to install packages offline. Check it out! https://www.kaggle.com/code/analokamus/how-to-use-pyvips-offline/notebook", "1857479": "Good idea, this could be useful to many of us while submitting our work!", "1868040": "I wrote a blog about it how you can use one notebook as a dataset for another kernel that is running offline:\nhttps://towardsdatascience.com/easy-kaggle-offline-submission-with-chaining-kernels-30bba5ea5c4d", "1858308": "This is a great tip! I had the same issue with trying to load some pretrained weights. I ended up just downloading the files and compressing them as ZIP file, uploading it to my Kaggle account, and then importing it like you said. Worked like a charm!\n", "1872152": "" } }