{
  "id": 316870,
  "title": "Tensorflow recommenders, need some help",
  "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/316870",
  "author_name": "",
  "post_date": "2022-04-04T10:58:52.744877600Z",
  "votes": 4,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Has anyone gotten TFRS working with the dataset from this contest? I've been struggling with it in the below notebook, but as of now I'm stuck on just not getting it into a format that it likes. I got it into an Eager Tensor, which I'm not sure how to get it into a mapped tensor, which I think is what it needs. I'm hoping that someone could point me in the right direction, there doesn't seem to be any examples of TFRS being used with data coming from a custom pandas/cudf dataframe.</p>\n<p>My notebook is here: <br>\n<a href=\"https://www.kaggle.com/code/beezus666/tensor-flow-recommender-quick-start?scriptVersionId=92067251\" target=\"_blank\">https://www.kaggle.com/code/beezus666/tensor-flow-recommender-quick-start?scriptVersionId=92067251</a></p>\n<p>Thanks!</p>",
  "messages": [
    {
      "id": "1744827",
      "postDate": "04/04/2022 10:58:52",
      "content": "<p>Has anyone gotten TFRS working with the dataset from this contest? I've been struggling with it in the below notebook, but as of now I'm stuck on just not getting it into a format that it likes. I got it into an Eager Tensor, which I'm not sure how to get it into a mapped tensor, which I think is what it needs. I'm hoping that someone could point me in the right direction, there doesn't seem to be any examples of TFRS being used with data coming from a custom pandas/cudf dataframe.</p>\n<p>My notebook is here: <br>\n<a href=\"https://www.kaggle.com/code/beezus666/tensor-flow-recommender-quick-start?scriptVersionId=92067251\" target=\"_blank\">https://www.kaggle.com/code/beezus666/tensor-flow-recommender-quick-start?scriptVersionId=92067251</a></p>\n<p>Thanks!</p>",
      "rawMarkdown": "Has anyone gotten TFRS working with the dataset from this contest? I've been struggling with it in the below notebook, but as of now I'm stuck on just not getting it into a format that it likes. I got it into an Eager Tensor, which I'm not sure how to get it into a mapped tensor, which I think is what it needs. I'm hoping that someone could point me in the right direction, there doesn't seem to be any examples of TFRS being used with data coming from a custom pandas/cudf dataframe.\n\nMy notebook is here: \nhttps://www.kaggle.com/code/beezus666/tensor-flow-recommender-quick-start?scriptVersionId=92067251\n\nThanks!",
      "votes": null
    },
    {
      "id": "1746517",
      "postDate": "04/05/2022 22:08:40",
      "content": "<p>This was one of the first approaches that I attempted to do as well. But installing TFRS creates dependency issues.  I got it working once, by trying a TFRS version compatible with TF version 2.6. But I don't remember the exact version, it can be checked online :( HOWEVER,  and this is when I just gave up, after I managed to get it working, the installed packages created conflicts with the GPUs (cuda stuff that I didn't quite get), and everythinbg was being calculated on CPU, which made everything painfully slow).</p>\n<p>Maybe you could try using Google Colab. I didn't because I was too lazy to transfer all the data to my Google Drive. But it should work fine over there.</p>",
      "rawMarkdown": "This was one of the first approaches that I attempted to do as well. But installing TFRS creates dependency issues.  I got it working once, by trying a TFRS version compatible with TF version 2.6. But I don't remember the exact version, it can be checked online :( HOWEVER,  and this is when I just gave up, after I managed to get it working, the installed packages created conflicts with the GPUs (cuda stuff that I didn't quite get), and everythinbg was being calculated on CPU, which made everything painfully slow).\n\nMaybe you could try using Google Colab. I didn't because I was too lazy to transfer all the data to my Google Drive. But it should work fine over there.",
      "votes": null
    },
    {
      "id": "1746610",
      "postDate": "04/06/2022 00:35:42",
      "content": "<p>Thanks! I was coming around to thinking that I'd just give up on it and do something else… </p>\n<p>Can you share your TFRS notebook? I'd like to see where I went wrong.</p>",
      "rawMarkdown": "Thanks! I was coming around to thinking that I'd just give up on it and do something else... \n\nCan you share your TFRS notebook? I'd like to see where I went wrong.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1746517,
      "author_name": "patabu2",
      "author_url": "",
      "post_date": "04/05/2022 22:08:40",
      "content": "<p>This was one of the first approaches that I attempted to do as well. But installing TFRS creates dependency issues.  I got it working once, by trying a TFRS version compatible with TF version 2.6. But I don't remember the exact version, it can be checked online :( HOWEVER,  and this is when I just gave up, after I managed to get it working, the installed packages created conflicts with the GPUs (cuda stuff that I didn't quite get), and everythinbg was being calculated on CPU, which made everything painfully slow).</p>\n<p>Maybe you could try using Google Colab. I didn't because I was too lazy to transfer all the data to my Google Drive. But it should work fine over there.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1746610,
          "author_name": "beezus666",
          "author_url": "",
          "post_date": "04/06/2022 00:35:42",
          "content": "<p>Thanks! I was coming around to thinking that I'd just give up on it and do something else… </p>\n<p>Can you share your TFRS notebook? I'd like to see where I went wrong.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1744827": "Has anyone gotten TFRS working with the dataset from this contest? I've been struggling with it in the below notebook, but as of now I'm stuck on just not getting it into a format that it likes. I got it into an Eager Tensor, which I'm not sure how to get it into a mapped tensor, which I think is what it needs. I'm hoping that someone could point me in the right direction, there doesn't seem to be any examples of TFRS being used with data coming from a custom pandas/cudf dataframe.\n\nMy notebook is here: \nhttps://www.kaggle.com/code/beezus666/tensor-flow-recommender-quick-start?scriptVersionId=92067251\n\nThanks!",
    "1746517": "This was one of the first approaches that I attempted to do as well. But installing TFRS creates dependency issues.  I got it working once, by trying a TFRS version compatible with TF version 2.6. But I don't remember the exact version, it can be checked online :( HOWEVER,  and this is when I just gave up, after I managed to get it working, the installed packages created conflicts with the GPUs (cuda stuff that I didn't quite get), and everythinbg was being calculated on CPU, which made everything painfully slow).\n\nMaybe you could try using Google Colab. I didn't because I was too lazy to transfer all the data to my Google Drive. But it should work fine over there.",
    "1746610": "Thanks! I was coming around to thinking that I'd just give up on it and do something else... \n\nCan you share your TFRS notebook? I'd like to see where I went wrong."
  },
  "source": "meta"
}