{
  "id": 453126,
  "title": "Can you provide evalution code?",
  "url": "/competitions/audio-driven-talking-head-generation/discussion/453126",
  "author_name": "",
  "post_date": "2023-11-05T03:36:50.550671Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I'm interested in the single image setting of this task, can you provide the evaluation code (FID, lip-sync) so that we can evaluate ourselves after the model training?</p>",
  "messages": [
    {
      "id": "2512939",
      "postDate": "11/05/2023 03:36:50",
      "content": "<p>I'm interested in the single image setting of this task, can you provide the evaluation code (FID, lip-sync) so that we can evaluate ourselves after the model training?</p>",
      "rawMarkdown": "I'm interested in the single image setting of this task, can you provide the evaluation code (FID, lip-sync) so that we can evaluate ourselves after the model training?",
      "votes": null
    },
    {
      "id": "2513297",
      "postDate": "11/05/2023 11:17:11",
      "content": "<p>Thank you for your interest! The complete evaluation code would be provided after the deadline of the competition. However, I could provide the code base of our evaluation script. We use <strong><a href=\"https://github.com/joonson/syncnet_python\" target=\"_blank\">syncnet</a></strong> to evaluate the lip synchronization and use FID from the \"<strong>piq</strong>\" package. You could consider implement yourself. </p>",
      "rawMarkdown": "Thank you for your interest! The complete evaluation code would be provided after the deadline of the competition. However, I could provide the code base of our evaluation script. We use **[syncnet](https://github.com/joonson/syncnet_python)** to evaluate the lip synchronization and use FID from the \"**piq**\" package. You could consider implement yourself.",
      "votes": null
    },
    {
      "id": "2513451",
      "postDate": "11/05/2023 13:09:54",
      "content": "<p>okay! thanks for replying~</p>",
      "rawMarkdown": "okay! thanks for replying~",
      "votes": null
    },
    {
      "id": "2514311",
      "postDate": "11/06/2023 07:24:28",
      "content": "<p>You're welcome.</p>",
      "rawMarkdown": "You're welcome.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2513297,
      "author_name": "jingnangao",
      "author_url": "",
      "post_date": "11/05/2023 11:17:11",
      "content": "<p>Thank you for your interest! The complete evaluation code would be provided after the deadline of the competition. However, I could provide the code base of our evaluation script. We use <strong><a href=\"https://github.com/joonson/syncnet_python\" target=\"_blank\">syncnet</a></strong> to evaluate the lip synchronization and use FID from the \"<strong>piq</strong>\" package. You could consider implement yourself. </p>",
      "votes": null,
      "replies": [
        {
          "id": 2513451,
          "author_name": "sunxusen",
          "author_url": "",
          "post_date": "11/05/2023 13:09:54",
          "content": "<p>okay! thanks for replying~</p>",
          "votes": null,
          "replies": [
            {
              "id": 2514311,
              "author_name": "jingnangao",
              "author_url": "",
              "post_date": "11/06/2023 07:24:28",
              "content": "<p>You're welcome.</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2512939": "I'm interested in the single image setting of this task, can you provide the evaluation code (FID, lip-sync) so that we can evaluate ourselves after the model training?",
    "2513297": "Thank you for your interest! The complete evaluation code would be provided after the deadline of the competition. However, I could provide the code base of our evaluation script. We use **[syncnet](https://github.com/joonson/syncnet_python)** to evaluate the lip synchronization and use FID from the \"**piq**\" package. You could consider implement yourself.",
    "2513451": "okay! thanks for replying~",
    "2514311": "You're welcome."
  },
  "source": "meta"
}