{ "id": 163303, "title": "External packages not allowed?", "url": "/competitions/siim-isic-melanoma-classification/discussion/163303", "author_name": "NAIN", "post_date": "2020-07-01T15:43:09.504000", "votes": 2, "comment_count": 8, "views": 0, "content": "

I am trying to upgrade the TF version in my environment. Even though I uninstalled and installed tf-nightly, it still picks up 2.2.

\n\n

\"\"

", "messages": [ { "id": 911164, "postDate": "2020-07-01T15:43:09.503Z", "content": "

I am trying to upgrade the TF version in my environment. Even though I uninstalled and installed tf-nightly, it still picks up 2.2.

\n\n

\"\"

", "rawMarkdown": "I am trying to upgrade the TF version in my environment. Even though I uninstalled and installed tf-nightly, it still picks up `2.2`. \n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F536977%2F358e0138d532064e2af5ed7b56a52a62%2FScreenshot%202020-07-01%20at%209.10.54%20PM.png?generation=1593618178459248&alt=media)\n", "votes": 2 }, { "id": 911331, "postDate": "2020-07-01T16:59:47.043Z", "content": "

If GPU is enabled, you should uninstall tensorflow-gpu and install tf-nightly-gpu

", "rawMarkdown": "If GPU is enabled, you should uninstall **tensorflow-gpu** and install **tf-nightly-gpu**", "replies": [ { "id": 911389, "postDate": "2020-07-01T17:40:08.050Z", "content": "

@sirishks does it make any difference?? \nI mean in gpu perfomances?

", "rawMarkdown": "@sirishks does it make any difference?? \nI mean in gpu perfomances?" }, { "id": 911487, "postDate": "2020-07-01T18:59:44.403Z", "content": "

There is no separate binary for GPU now

", "rawMarkdown": "There is no separate binary for GPU now" } ] }, { "id": 911292, "postDate": "2020-07-01T16:40:17.583Z", "content": "

The best way tto deal with that kind of problem is to work locally.\nIn your command prompt runned as administrator you can install and update your packages.\nThere are some problem with kaggle notebook sometimes and there is a risk to lost your work.

", "rawMarkdown": "The best way tto deal with that kind of problem is to work locally.\nIn your command prompt runned as administrator you can install and update your packages.\nThere are some problem with kaggle notebook sometimes and there is a risk to lost your work." }, { "id": 911286, "postDate": "2020-07-01T16:36:31.640Z", "content": "

If all you need is Efficientnet, consider:

\n\n
!pip install -q efficientnet\n
", "rawMarkdown": "If all you need is Efficientnet, consider:\n\n !pip install -q efficientnet\n" }, { "id": 911768, "postDate": "2020-07-02T02:32:43.520Z", "rawMarkdown": "", "votes": 3, "isDeleted": true, "replies": [ { "id": 916359, "postDate": "2020-07-05T15:23:59.673Z", "content": "

You are the mvp :)

", "rawMarkdown": "You are the mvp :)" }, { "id": 916383, "postDate": "2020-07-05T15:38:27.897Z", "rawMarkdown": "", "isDeleted": true } ] } ], "comments": [ { "id": 911331, "author_name": "Sirish Somanchi", "author_url": "", "post_date": "2020-07-01T16:59:47.043000", "content": "

If GPU is enabled, you should uninstall tensorflow-gpu and install tf-nightly-gpu

", "votes": 0, "replies": [ { "id": 911389, "author_name": "MhdSharuk", "author_url": "", "post_date": "2020-07-01T17:40:08.050000", "content": "

@sirishks does it make any difference?? \nI mean in gpu perfomances?

", "votes": 0, "replies": [] }, { "id": 911487, "author_name": "NAIN", "author_url": "", "post_date": "2020-07-01T18:59:44.403000", "content": "

There is no separate binary for GPU now

", "votes": 0, "replies": [] } ] }, { "id": 911292, "author_name": "Koara", "author_url": "", "post_date": "2020-07-01T16:40:17.583000", "content": "

The best way tto deal with that kind of problem is to work locally.\nIn your command prompt runned as administrator you can install and update your packages.\nThere are some problem with kaggle notebook sometimes and there is a risk to lost your work.

", "votes": 0, "replies": [] }, { "id": 911286, "author_name": "quadcore/Richard Epstein", "author_url": "", "post_date": "2020-07-01T16:36:31.640000", "content": "

If all you need is Efficientnet, consider:

\n\n
!pip install -q efficientnet\n
", "votes": 0, "replies": [] }, { "id": 911768, "author_name": "", "author_url": "", "post_date": "2020-07-02T02:32:43.520000", "content": "", "votes": 3, "replies": [ { "id": 916359, "author_name": "Alin Cijov", "author_url": "", "post_date": "2020-07-05T15:23:59.673000", "content": "

You are the mvp :)

", "votes": 0, "replies": [] }, { "id": 916383, "author_name": "", "author_url": "", "post_date": "2020-07-05T15:38:27.897000", "content": "", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "911164": "I am trying to upgrade the TF version in my environment. Even though I uninstalled and installed tf-nightly, it still picks up `2.2`. \n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F536977%2F358e0138d532064e2af5ed7b56a52a62%2FScreenshot%202020-07-01%20at%209.10.54%20PM.png?generation=1593618178459248&alt=media)\n", "911331": "If GPU is enabled, you should uninstall **tensorflow-gpu** and install **tf-nightly-gpu**", "911292": "The best way tto deal with that kind of problem is to work locally.\nIn your command prompt runned as administrator you can install and update your packages.\nThere are some problem with kaggle notebook sometimes and there is a risk to lost your work.", "911286": "If all you need is Efficientnet, consider:\n\n !pip install -q efficientnet\n", "911768": "" } }