{
  "id": 433110,
  "title": "Useful External Datasets, Models and Resources",
  "url": "/competitions/bengaliai-speech/discussion/433110",
  "author_name": "",
  "post_date": "2023-08-20T10:26:33.001852300Z",
  "votes": 17,
  "comment_count": 7,
  "views": 0,
  "content": "<p><strong>Kaggle past related competitions:</strong></p>\n<p>homepage:  <a href=\"https://www.kaggle.com/competitions/dlsprint/\" target=\"_blank\">https://www.kaggle.com/competitions/dlsprint/</a></p>\n<ul>\n<li><p>1st place solution's training notebook:<br>\n<a href=\"https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2</a></p></li>\n<li><p>1st place solution's inference notebook:<br>\n<a href=\"https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2</a></p></li>\n</ul>\n<p><strong>Hugging Face public models:</strong></p>\n<ul>\n<li><p><a href=\"https://huggingface.co/arijitx/wav2vec2-xls-r-300m-bengali\" target=\"_blank\">https://huggingface.co/arijitx/wav2vec2-xls-r-300m-bengali</a></p></li>\n<li><p><a href=\"https://huggingface.co/ai4bharat/indicwav2vec_v1_bengali\" target=\"_blank\">https://huggingface.co/ai4bharat/indicwav2vec_v1_bengali</a></p></li>\n<li><p><a href=\"https://huggingface.co/bengaliAI/BanglaConformer\" target=\"_blank\">https://huggingface.co/bengaliAI/BanglaConformer</a></p></li>\n<li><p><a href=\"https://huggingface.co/arijitx/wav2vec2-large-xlsr-bengali\" target=\"_blank\">https://huggingface.co/arijitx/wav2vec2-large-xlsr-bengali</a></p></li>\n<li><p><a href=\"https://huggingface.co/shahruk10/wav2vec2-xls-r-300m-bengali-commonvoice\" target=\"_blank\">https://huggingface.co/shahruk10/wav2vec2-xls-r-300m-bengali-commonvoice</a></p></li>\n<li><p><a href=\"https://huggingface.co/tanmoyio/wav2vec2-large-xlsr-bengali\" target=\"_blank\">https://huggingface.co/tanmoyio/wav2vec2-large-xlsr-bengali</a></p></li>\n</ul>\n<p><strong>Extra ASR datasets:</strong></p>\n<ul>\n<li><p><a href=\"https://huggingface.co/datasets/parambharat/bengali_asr_corpus\" target=\"_blank\">https://huggingface.co/datasets/parambharat/bengali_asr_corpus</a></p></li>\n<li><p><a href=\"https://www.kaggle.com/datasets/keshan/large-bengali-asr-dataset\" target=\"_blank\">https://www.kaggle.com/datasets/keshan/large-bengali-asr-dataset</a></p></li>\n<li><p><a href=\"https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k\" target=\"_blank\">https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k</a></p></li>\n</ul>",
  "messages": [
    {
      "id": "2399413",
      "postDate": "08/20/2023 10:26:33",
      "content": "<p><strong>Kaggle past related competitions:</strong></p>\n<p>homepage:  <a href=\"https://www.kaggle.com/competitions/dlsprint/\" target=\"_blank\">https://www.kaggle.com/competitions/dlsprint/</a></p>\n<ul>\n<li><p>1st place solution's training notebook:<br>\n<a href=\"https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2</a></p></li>\n<li><p>1st place solution's inference notebook:<br>\n<a href=\"https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2</a></p></li>\n</ul>\n<p><strong>Hugging Face public models:</strong></p>\n<ul>\n<li><p><a href=\"https://huggingface.co/arijitx/wav2vec2-xls-r-300m-bengali\" target=\"_blank\">https://huggingface.co/arijitx/wav2vec2-xls-r-300m-bengali</a></p></li>\n<li><p><a href=\"https://huggingface.co/ai4bharat/indicwav2vec_v1_bengali\" target=\"_blank\">https://huggingface.co/ai4bharat/indicwav2vec_v1_bengali</a></p></li>\n<li><p><a href=\"https://huggingface.co/bengaliAI/BanglaConformer\" target=\"_blank\">https://huggingface.co/bengaliAI/BanglaConformer</a></p></li>\n<li><p><a href=\"https://huggingface.co/arijitx/wav2vec2-large-xlsr-bengali\" target=\"_blank\">https://huggingface.co/arijitx/wav2vec2-large-xlsr-bengali</a></p></li>\n<li><p><a href=\"https://huggingface.co/shahruk10/wav2vec2-xls-r-300m-bengali-commonvoice\" target=\"_blank\">https://huggingface.co/shahruk10/wav2vec2-xls-r-300m-bengali-commonvoice</a></p></li>\n<li><p><a href=\"https://huggingface.co/tanmoyio/wav2vec2-large-xlsr-bengali\" target=\"_blank\">https://huggingface.co/tanmoyio/wav2vec2-large-xlsr-bengali</a></p></li>\n</ul>\n<p><strong>Extra ASR datasets:</strong></p>\n<ul>\n<li><p><a href=\"https://huggingface.co/datasets/parambharat/bengali_asr_corpus\" target=\"_blank\">https://huggingface.co/datasets/parambharat/bengali_asr_corpus</a></p></li>\n<li><p><a href=\"https://www.kaggle.com/datasets/keshan/large-bengali-asr-dataset\" target=\"_blank\">https://www.kaggle.com/datasets/keshan/large-bengali-asr-dataset</a></p></li>\n<li><p><a href=\"https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k\" target=\"_blank\">https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k</a></p></li>\n</ul>",
      "rawMarkdown": "**Kaggle past related competitions:**\n\nhomepage:  https://www.kaggle.com/competitions/dlsprint/\n\n- 1st place solution's training notebook:\nhttps://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\n\n- 1st place solution's inference notebook:\nhttps://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\n\n**Hugging Face public models:**\n\n- https://huggingface.co/arijitx/wav2vec2-xls-r-300m-bengali\n\n- https://huggingface.co/ai4bharat/indicwav2vec_v1_bengali\n\n- https://huggingface.co/bengaliAI/BanglaConformer\n\n- https://huggingface.co/arijitx/wav2vec2-large-xlsr-bengali\n\n- https://huggingface.co/shahruk10/wav2vec2-xls-r-300m-bengali-commonvoice\n\n- https://huggingface.co/tanmoyio/wav2vec2-large-xlsr-bengali\n\n**Extra ASR datasets:**\n\n- https://huggingface.co/datasets/parambharat/bengali_asr_corpus\n\n- https://www.kaggle.com/datasets/keshan/large-bengali-asr-dataset\n\n- https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k",
      "votes": null
    },
    {
      "id": "2399434",
      "postDate": "08/20/2023 10:44:25",
      "content": "<p><a href=\"https://www.kaggle.com/berserker408\" target=\"_blank\">@berserker408</a> great resources, thanks for sharing. </p>",
      "rawMarkdown": "berserker408 great resources, thanks for sharing.",
      "votes": null
    },
    {
      "id": "2400166",
      "postDate": "08/20/2023 20:59:30",
      "content": "<p>Thanks for sharing! Very informative</p>",
      "rawMarkdown": "Thanks for sharing! Very informative",
      "votes": null
    },
    {
      "id": "2401009",
      "postDate": "08/21/2023 11:33:40",
      "content": "<p><a href=\"https://www.kaggle.com/berserker408\" target=\"_blank\">@berserker408</a> fabulous contributions to the new-comer, wish you have a satisfactory ranking in this competition.</p>",
      "rawMarkdown": "berserker408 fabulous contributions to the new-comer, wish you have a satisfactory ranking in this competition.",
      "votes": null
    },
    {
      "id": "2407363",
      "postDate": "08/25/2023 03:50:55",
      "content": "<blockquote>\n  <p>1st place solution's training notebook:<br>\n  <a href=\"https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2</a></p>\n  <p>1st place solution's inference notebook:<br>\n  <a href=\"https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2</a></p>\n</blockquote>\n<p>These two notebooks are not available now  ：(</p>",
      "rawMarkdown": ">1st place solution's training notebook:\nhttps://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\n\n>1st place solution's inference notebook:\nhttps://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\n\nThese two notebooks are not available now  ：(",
      "votes": null
    },
    {
      "id": "2407682",
      "postDate": "08/25/2023 07:43:08",
      "content": "<p>Isn't it the same as this one ? <br>\n<a href=\"https://www.kaggle.com/code/mbmmurad/lb-0-49-wav2vec2-baseline-train-and-infer?scriptVersionId=139805745\" target=\"_blank\">https://www.kaggle.com/code/mbmmurad/lb-0-49-wav2vec2-baseline-train-and-infer?scriptVersionId=139805745</a><br>\nThe first version has training code</p>",
      "rawMarkdown": "Isn't it the same as this one ? \nhttps://www.kaggle.com/code/mbmmurad/lb-0-49-wav2vec2-baseline-train-and-infer?scriptVersionId=139805745\nThe first version has training code",
      "votes": null
    },
    {
      "id": "2408091",
      "postDate": "08/25/2023 12:39:15",
      "content": "<p>Yes it is. So sad…</p>",
      "rawMarkdown": "Yes it is. So sad...",
      "votes": null
    },
    {
      "id": "2410443",
      "postDate": "08/27/2023 00:41:22",
      "content": "<p><a href=\"https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k\" target=\"_blank\">https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k</a></p>\n<p>this is just a 10k sample of the dataset from this competition by the way. just preprocessed for whisper</p>",
      "rawMarkdown": "https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k\n\nthis is just a 10k sample of the dataset from this competition by the way. just preprocessed for whisper",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2399434,
      "author_name": "faysalmiah1721758",
      "author_url": "",
      "post_date": "08/20/2023 10:44:25",
      "content": "<p><a href=\"https://www.kaggle.com/berserker408\" target=\"_blank\">@berserker408</a> great resources, thanks for sharing. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2400166,
      "author_name": "suezkingfml",
      "author_url": "",
      "post_date": "08/20/2023 20:59:30",
      "content": "<p>Thanks for sharing! Very informative</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2401009,
      "author_name": "liushundawang",
      "author_url": "",
      "post_date": "08/21/2023 11:33:40",
      "content": "<p><a href=\"https://www.kaggle.com/berserker408\" target=\"_blank\">@berserker408</a> fabulous contributions to the new-comer, wish you have a satisfactory ranking in this competition.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2407363,
      "author_name": "qiaoshiji",
      "author_url": "",
      "post_date": "08/25/2023 03:50:55",
      "content": "<blockquote>\n  <p>1st place solution's training notebook:<br>\n  <a href=\"https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2</a></p>\n  <p>1st place solution's inference notebook:<br>\n  <a href=\"https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\" target=\"_blank\">https://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2</a></p>\n</blockquote>\n<p>These two notebooks are not available now  ：(</p>",
      "votes": null,
      "replies": [
        {
          "id": 2407682,
          "author_name": "nyleve",
          "author_url": "",
          "post_date": "08/25/2023 07:43:08",
          "content": "<p>Isn't it the same as this one ? <br>\n<a href=\"https://www.kaggle.com/code/mbmmurad/lb-0-49-wav2vec2-baseline-train-and-infer?scriptVersionId=139805745\" target=\"_blank\">https://www.kaggle.com/code/mbmmurad/lb-0-49-wav2vec2-baseline-train-and-infer?scriptVersionId=139805745</a><br>\nThe first version has training code</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 2408091,
          "author_name": "berserker408",
          "author_url": "",
          "post_date": "08/25/2023 12:39:15",
          "content": "<p>Yes it is. So sad…</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2410443,
      "author_name": "nbroad",
      "author_url": "",
      "post_date": "08/27/2023 00:41:22",
      "content": "<p><a href=\"https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k\" target=\"_blank\">https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k</a></p>\n<p>this is just a 10k sample of the dataset from this competition by the way. just preprocessed for whisper</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2399413": "**Kaggle past related competitions:**\n\nhomepage:  https://www.kaggle.com/competitions/dlsprint/\n\n- 1st place solution's training notebook:\nhttps://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\n\n- 1st place solution's inference notebook:\nhttps://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\n\n**Hugging Face public models:**\n\n- https://huggingface.co/arijitx/wav2vec2-xls-r-300m-bengali\n\n- https://huggingface.co/ai4bharat/indicwav2vec_v1_bengali\n\n- https://huggingface.co/bengaliAI/BanglaConformer\n\n- https://huggingface.co/arijitx/wav2vec2-large-xlsr-bengali\n\n- https://huggingface.co/shahruk10/wav2vec2-xls-r-300m-bengali-commonvoice\n\n- https://huggingface.co/tanmoyio/wav2vec2-large-xlsr-bengali\n\n**Extra ASR datasets:**\n\n- https://huggingface.co/datasets/parambharat/bengali_asr_corpus\n\n- https://www.kaggle.com/datasets/keshan/large-bengali-asr-dataset\n\n- https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k",
    "2399434": "berserker408 great resources, thanks for sharing.",
    "2400166": "Thanks for sharing! Very informative",
    "2401009": "berserker408 fabulous contributions to the new-comer, wish you have a satisfactory ranking in this competition.",
    "2407363": ">1st place solution's training notebook:\nhttps://www.kaggle.com/code/shahruk10/training-notebook-wav2vec2\n\n>1st place solution's inference notebook:\nhttps://www.kaggle.com/code/shahruk10/inference-notebook-wav2vec2\n\nThese two notebooks are not available now  ：(",
    "2407682": "Isn't it the same as this one ? \nhttps://www.kaggle.com/code/mbmmurad/lb-0-49-wav2vec2-baseline-train-and-infer?scriptVersionId=139805745\nThe first version has training code",
    "2408091": "Yes it is. So sad...",
    "2410443": "https://www.kaggle.com/datasets/nbroad/bengali-ai-asr-10k\n\nthis is just a 10k sample of the dataset from this competition by the way. just preprocessed for whisper"
  },
  "source": "meta"
}