{
  "id": 357684,
  "title": " River Tool: Online Learning for ML Models in Production",
  "url": "/competitions/tabular-playground-series-oct-2022/discussion/357684",
  "author_name": "Rashmi Margani",
  "post_date": "2022-10-05T08:48:45.565000",
  "votes": 12,
  "comment_count": 0,
  "views": 0,
  "content": "<p><strong>One of the MOST FRUSTRATING aspects of running machine learning models in the real world?</strong></p>\n<p>You have to keep updating and retraining your ML models over time, and this becomes very difficult in scenarios when you need to automate this.</p>\n<p>This is where online learning comes into play, where you can keep training ML models on incremental and streaming data automatically! Check out the <a href=\"https://riverml.xyz/0.13.0/\" target=\"_blank\">River</a> open-source library which enables you to do this with ease in just a few lines of code as shown in the example below!</p>\n<p>Default sci-kit-learn models have this support only for a few restrictive models and this is where River builds on this and has support for a vast number of supervised and unsupervised models along with other aspects like drift detection.</p>\n<p><strong>Thanks &amp; if you find this content useful &amp; to create relative content like this, then please click the \"upvote\" button in the top-right -- it's very encouraging to know when people appreciate the work &amp; support it .¶</strong></p>",
  "messages": [
    {
      "id": 1972619,
      "postDate": "2022-10-05T08:48:45.567Z",
      "content": "<p><strong>One of the MOST FRUSTRATING aspects of running machine learning models in the real world?</strong></p>\n<p>You have to keep updating and retraining your ML models over time, and this becomes very difficult in scenarios when you need to automate this.</p>\n<p>This is where online learning comes into play, where you can keep training ML models on incremental and streaming data automatically! Check out the <a href=\"https://riverml.xyz/0.13.0/\" target=\"_blank\">River</a> open-source library which enables you to do this with ease in just a few lines of code as shown in the example below!</p>\n<p>Default sci-kit-learn models have this support only for a few restrictive models and this is where River builds on this and has support for a vast number of supervised and unsupervised models along with other aspects like drift detection.</p>\n<p><strong>Thanks &amp; if you find this content useful &amp; to create relative content like this, then please click the \"upvote\" button in the top-right -- it's very encouraging to know when people appreciate the work &amp; support it .¶</strong></p>",
      "rawMarkdown": "**One of the MOST FRUSTRATING aspects of running machine learning models in the real world?**\n\nYou have to keep updating and retraining your ML models over time, and this becomes very difficult in scenarios when you need to automate this.\n\nThis is where online learning comes into play, where you can keep training ML models on incremental and streaming data automatically! Check out the [River](https://riverml.xyz/0.13.0/) open-source library which enables you to do this with ease in just a few lines of code as shown in the example below!\n\nDefault sci-kit-learn models have this support only for a few restrictive models and this is where River builds on this and has support for a vast number of supervised and unsupervised models along with other aspects like drift detection.\n\n**Thanks & if you find this content useful & to create relative content like this, then please click the \"upvote\" button in the top-right -- it's very encouraging to know when people appreciate the work & support it .¶**",
      "votes": 12
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1972619": "**One of the MOST FRUSTRATING aspects of running machine learning models in the real world?**\n\nYou have to keep updating and retraining your ML models over time, and this becomes very difficult in scenarios when you need to automate this.\n\nThis is where online learning comes into play, where you can keep training ML models on incremental and streaming data automatically! Check out the [River](https://riverml.xyz/0.13.0/) open-source library which enables you to do this with ease in just a few lines of code as shown in the example below!\n\nDefault sci-kit-learn models have this support only for a few restrictive models and this is where River builds on this and has support for a vast number of supervised and unsupervised models along with other aspects like drift detection.\n\n**Thanks & if you find this content useful & to create relative content like this, then please click the \"upvote\" button in the top-right -- it's very encouraging to know when people appreciate the work & support it .¶**"
  }
}