The competition rule says:
\n Source: You hereby license will license your winning Submission the source code used generate the Submission under an Source Initiative-approved license (see www.opensource.org) that event limits commercial use such code model containing depending such code. the extent your Submission makes use generally commercially available software you that you used generate your submission, but that can be procured the Competition Sponsor undue expense, you the license the sentence that software.\nSo it clearly forbids models and datasets e.g. under CC-BY-NC such facebook/mms-1b.
\nHere are list of Bengali datasets which we could use:
\n@imtiazprio Could you confirm whether we can use RESPIN/MADASR? There is no licence info available.
", "messages": [ { "id": "2439953", "postDate": "09/15/2023 07:23:26", "content": "The competition rule says:
\n Source: You hereby license will license your winning Submission the source code used generate the Submission under an Source Initiative-approved license (see www.opensource.org) that event limits commercial use such code model containing depending such code. the extent your Submission makes use generally commercially available software you that you used generate your submission, but that can be procured the Competition Sponsor undue expense, you the license the sentence that software.\nSo it clearly forbids models and datasets e.g. under CC-BY-NC such facebook/mms-1b.
\nHere are list of Bengali datasets which we could use:
\n@imtiazprio Could you confirm whether we can use RESPIN/MADASR? There is no licence info available.
", "rawMarkdown": "The competition rule says:\n```\nOpen Source: You hereby license and will license your winning Submission and the source code used to generate the Submission under an Open Source Initiative-approved license (see www.opensource.org) that in no event limits commercial use of such code or model containing or depending on such code. To the extent your Submission makes use of generally commercially available software not owned by you that you used to generate your submission, but that can be procured by the Competition Sponsor without undue expense, you do not grant the license in the preceding sentence to that software.\n```\n\nSo it clearly forbids models and datasets e.g. under CC-BY-NC such facebook/mms-1b.\n\nHere are list of Bengali datasets which we could use:\n- https://www.openslr.org/53/ CC BY-SA 4.0 -> should be OK\n- https://www.openslr.org/37/ CC BY-SA 4.0 -> should be OK\n- https://ai4bharat.iitm.ac.in/shrutilipi/ CC0 1.0 -> should be OK\n- https://www.iitm.ac.in/donlab/tts/ -> CC BY-SA 2.0 -> should be OK\n- https://sites.google.com/view/respinasrchallenge2023/dataset (RESPIN/MADASR) -> no license info\n\n@imtiazprio Could you confirm whether we can use RESPIN/MADASR? There is no licence info available.", "votes": null }, { "id": "2439958", "postDate": "09/15/2023 07:27:32", "content": "Does someone have access to https://github.com/Open-Speech-EkStep/ULCA-asr-dataset-corpus (CC BY-SA 4.0)?
", "rawMarkdown": "Does someone have access to https://github.com/Open-Speech-EkStep/ULCA-asr-dataset-corpus (CC BY-SA 4.0)?", "votes": null }, { "id": "2440417", "postDate": "09/15/2023 13:42:25", "content": "Will we be removed from the Leaderboard if we use facebook/mms-1b ?
", "rawMarkdown": "Will we be removed from the Leaderboard if we use facebook/mms-1b ?", "votes": null }, { "id": "2440548", "postDate": "09/15/2023 15:17:25", "content": "@tugstugi The indic-TTS dataset isn't under the CC BY-SA 2.0. Maybe you thought it because of the text on their website: \"Left photo credit: Indic Scripts, 2013, Licensed under CC BY-SA 2.0\". They have a custom license for the dataset that you can find here. But in the 2.2 section of the license, it seems that the dataset has some restrictions on how you can explore commercially any derivative work (like a model). And the only place I find a person who maintains the dataset talking about license says that \"you can download the data for research purposes\". So, just in case, I've decided not to use that dataset in my model training, but if one of the competition hosts, like @imtiazprio, says that it's ok to use the indic-TTS, I'll probably add it to my training data.
\nAbout the RESPIN/MADASR, it's ok to use that dataset. It's under the CC BY 4.0. You can check its license here
", "rawMarkdown": "tugstugi The indic-TTS dataset isn't under the CC BY-SA 2.0. Maybe you thought it because of the text on their [website](https://www.iitm.ac.in/donlab/tts/index.php): \"Left photo credit: Indic Scripts, 2013, Licensed under CC BY-SA 2.0\". They have a custom license for the dataset that you can find [here](https://www.iitm.ac.in/donlab/tts/synthDocs.php). But in the 2.2 section of the license, it seems that the dataset has some restrictions on how you can explore commercially any derivative work (like a model). And the only [place](https://groups.google.com/g/indictts/c/uJ7Bz6glwFo) I find a person who maintains the dataset talking about license says that \"you can download the data for research purposes\". So, just in case, I've decided not to use that dataset in my model training, but if one of the competition hosts, like @imtiazprio, says that it's ok to use the indic-TTS, I'll probably add it to my training data.\n\nAbout the RESPIN/MADASR, it's ok to use that dataset. It's under the CC BY 4.0. You can check its license [here](https://ee.iisc.ac.in/madasr23dataset/RESPIN_Data_TAC.pdf)", "votes": null }, { "id": "2440555", "postDate": "09/15/2023 15:19:52", "content": "thanks @jonatasgrosman
", "rawMarkdown": "thanks @jonatasgrosman", "votes": null } ], "comments": [ { "id": 2439958, "author_name": "tugstugi", "author_url": "", "post_date": "09/15/2023 07:27:32", "content": "Does someone have access to https://github.com/Open-Speech-EkStep/ULCA-asr-dataset-corpus (CC BY-SA 4.0)?
", "votes": null, "replies": [] }, { "id": 2440417, "author_name": "dhakshiin1601", "author_url": "", "post_date": "09/15/2023 13:42:25", "content": "Will we be removed from the Leaderboard if we use facebook/mms-1b ?
", "votes": null, "replies": [] }, { "id": 2440548, "author_name": "jonatasgrosman", "author_url": "", "post_date": "09/15/2023 15:17:25", "content": "@tugstugi The indic-TTS dataset isn't under the CC BY-SA 2.0. Maybe you thought it because of the text on their website: \"Left photo credit: Indic Scripts, 2013, Licensed under CC BY-SA 2.0\". They have a custom license for the dataset that you can find here. But in the 2.2 section of the license, it seems that the dataset has some restrictions on how you can explore commercially any derivative work (like a model). And the only place I find a person who maintains the dataset talking about license says that \"you can download the data for research purposes\". So, just in case, I've decided not to use that dataset in my model training, but if one of the competition hosts, like @imtiazprio, says that it's ok to use the indic-TTS, I'll probably add it to my training data.
\nAbout the RESPIN/MADASR, it's ok to use that dataset. It's under the CC BY 4.0. You can check its license here
", "votes": null, "replies": [ { "id": 2440555, "author_name": "tugstugi", "author_url": "", "post_date": "09/15/2023 15:19:52", "content": "thanks @jonatasgrosman
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "2439953": "The competition rule says:\n```\nOpen Source: You hereby license and will license your winning Submission and the source code used to generate the Submission under an Open Source Initiative-approved license (see www.opensource.org) that in no event limits commercial use of such code or model containing or depending on such code. To the extent your Submission makes use of generally commercially available software not owned by you that you used to generate your submission, but that can be procured by the Competition Sponsor without undue expense, you do not grant the license in the preceding sentence to that software.\n```\n\nSo it clearly forbids models and datasets e.g. under CC-BY-NC such facebook/mms-1b.\n\nHere are list of Bengali datasets which we could use:\n- https://www.openslr.org/53/ CC BY-SA 4.0 -> should be OK\n- https://www.openslr.org/37/ CC BY-SA 4.0 -> should be OK\n- https://ai4bharat.iitm.ac.in/shrutilipi/ CC0 1.0 -> should be OK\n- https://www.iitm.ac.in/donlab/tts/ -> CC BY-SA 2.0 -> should be OK\n- https://sites.google.com/view/respinasrchallenge2023/dataset (RESPIN/MADASR) -> no license info\n\n@imtiazprio Could you confirm whether we can use RESPIN/MADASR? There is no licence info available.", "2439958": "Does someone have access to https://github.com/Open-Speech-EkStep/ULCA-asr-dataset-corpus (CC BY-SA 4.0)?", "2440417": "Will we be removed from the Leaderboard if we use facebook/mms-1b ?", "2440548": "tugstugi The indic-TTS dataset isn't under the CC BY-SA 2.0. Maybe you thought it because of the text on their [website](https://www.iitm.ac.in/donlab/tts/index.php): \"Left photo credit: Indic Scripts, 2013, Licensed under CC BY-SA 2.0\". They have a custom license for the dataset that you can find [here](https://www.iitm.ac.in/donlab/tts/synthDocs.php). But in the 2.2 section of the license, it seems that the dataset has some restrictions on how you can explore commercially any derivative work (like a model). And the only [place](https://groups.google.com/g/indictts/c/uJ7Bz6glwFo) I find a person who maintains the dataset talking about license says that \"you can download the data for research purposes\". So, just in case, I've decided not to use that dataset in my model training, but if one of the competition hosts, like @imtiazprio, says that it's ok to use the indic-TTS, I'll probably add it to my training data.\n\nAbout the RESPIN/MADASR, it's ok to use that dataset. It's under the CC BY 4.0. You can check its license [here](https://ee.iisc.ac.in/madasr23dataset/RESPIN_Data_TAC.pdf)", "2440555": "thanks @jonatasgrosman" }, "source": "meta" }