{
  "id": 191700,
  "title": "Perf, A Cautionary Tale...",
  "url": "/competitions/cdp-unlocking-climate-solutions/discussion/191700",
  "author_name": "AlexFromIdaho",
  "post_date": "2020-10-18T03:47:29.500000",
  "votes": -1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>So I'm running spaCy for the first time today, and the resulting code has been running for 2 hours at least. I took a gander at TaskManager, and it looks like it's the RAM maxing out, not the CPU. I'm thinking about my object-model, and I think I'm running a vast number of spaCy Doc objects on intermediate lists in the iterative step with the initialization of my individual response entry class. I think I can probably get a big performance improvement by NOT initializing my slave object with Doc creation and instead running NLP on individual front-end queries. Will make attempts at seeing the big picture more difficult, however. Thoughts?</p>",
  "messages": [
    {
      "id": 1052639,
      "postDate": "2020-10-18T03:47:29.500Z",
      "content": "<p>So I'm running spaCy for the first time today, and the resulting code has been running for 2 hours at least. I took a gander at TaskManager, and it looks like it's the RAM maxing out, not the CPU. I'm thinking about my object-model, and I think I'm running a vast number of spaCy Doc objects on intermediate lists in the iterative step with the initialization of my individual response entry class. I think I can probably get a big performance improvement by NOT initializing my slave object with Doc creation and instead running NLP on individual front-end queries. Will make attempts at seeing the big picture more difficult, however. Thoughts?</p>",
      "rawMarkdown": "So I'm running spaCy for the first time today, and the resulting code has been running for 2 hours at least. I took a gander at TaskManager, and it looks like it's the RAM maxing out, not the CPU. I'm thinking about my object-model, and I think I'm running a vast number of spaCy Doc objects on intermediate lists in the iterative step with the initialization of my individual response entry class. I think I can probably get a big performance improvement by NOT initializing my slave object with Doc creation and instead running NLP on individual front-end queries. Will make attempts at seeing the big picture more difficult, however. Thoughts?",
      "votes": -1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1052639": "So I'm running spaCy for the first time today, and the resulting code has been running for 2 hours at least. I took a gander at TaskManager, and it looks like it's the RAM maxing out, not the CPU. I'm thinking about my object-model, and I think I'm running a vast number of spaCy Doc objects on intermediate lists in the iterative step with the initialization of my individual response entry class. I think I can probably get a big performance improvement by NOT initializing my slave object with Doc creation and instead running NLP on individual front-end queries. Will make attempts at seeing the big picture more difficult, however. Thoughts?"
  }
}