We have decided to release our code.
\nThere are 2 main reasons:
\nThese notebooks are copies of the ones used to generate the final submission as well as the final submission itself:
\nThis code has not been executed on Kaggle infrastructure and likely needs more resources.
", "messages": [ { "id": "2323987", "postDate": "06/30/2023 09:49:10", "content": "We have decided to release our code.
\nThere are 2 main reasons:
\nThese notebooks are copies of the ones used to generate the final submission as well as the final submission itself:
\nThis code has not been executed on Kaggle infrastructure and likely needs more resources.
", "rawMarkdown": "We have decided to release our code.\n\nThere are 2 main reasons:\n- sharing is Kaggle DNA and we want to give back,\n- this is a contribution to keep the ML an open field because we fear there is a threat with CloseAI.\n\nThese notebooks are copies of the ones used to generate the final submission as well as the final submission itself:\n- https://www.kaggle.com/code/pdnartreb/pspfgp-1st-place-gbdt-training,\n- https://www.kaggle.com/code/pdnartreb/pspfgp-1st-place-nn-pretraining,\n- https://www.kaggle.com/code/pdnartreb/pspfgp-1st-place-nn-training,\n- https://www.kaggle.com/code/pdnartreb/pspfgp-1st-place-inference.\n\n*This code has not been executed on Kaggle infrastructure and likely needs more resources.*", "votes": null }, { "id": "2324149", "postDate": "06/30/2023 12:18:06", "content": "I really appreciate your decision.
\nIt will be a very great source for people who want to dig into.
Amazing work, thanks for sharing this with us!
", "rawMarkdown": "Amazing work, thanks for sharing this with us!", "votes": null }, { "id": "2325500", "postDate": "07/01/2023 12:14:57", "content": "Thank you for sharing the code! It's great to see Kagglers contributing to the community by sharing their winning solutions. By making your code available, you not only give back to the community but also help in keeping machine learning an open field.
\nThe provided notebooks contain the code used to generate the final submission as well as the code for training and inference. It's worth noting that this code has not been executed on the Kaggle infrastructure and may require additional resources to run successfully.
\nOnce again, thank you for your contribution and for promoting openness in the field of machine learning!
", "rawMarkdown": "Thank you for sharing the code! It's great to see Kagglers contributing to the community by sharing their winning solutions. By making your code available, you not only give back to the community but also help in keeping machine learning an open field.\n\nThe provided notebooks contain the code used to generate the final submission as well as the code for training and inference. It's worth noting that this code has not been executed on the Kaggle infrastructure and may require additional resources to run successfully.\n\nOnce again, thank you for your contribution and for promoting openness in the field of machine learning!", "votes": null }, { "id": "2325657", "postDate": "07/01/2023 14:01:07", "content": "@pdnartreb Thank you for sharing your code!Well done on your 1st place achievement!
", "rawMarkdown": "pdnartreb Thank you for sharing your code!Well done on your 1st place achievement!", "votes": null }, { "id": "2330061", "postDate": "07/04/2023 17:29:32", "content": "Thanks for sharing learned a lot just reading through the code.
", "rawMarkdown": "Thanks for sharing learned a lot just reading through the code.", "votes": null }, { "id": "2335986", "postDate": "07/09/2023 02:47:03", "content": "@pdnartreb
\nThanks for sharing, and I really appreciate your decision.
\nI have learned a lot from your notebook. If possible, can you public your dataset?
\nnow, I cant reproduce your dataset
%%time\nMETADATA = pickle.load((, ))\ndataset = pickle.load((, ))\ntokenizer_map = pickle.load((, ))\npretrain_datasets = pickle.load((, ))\nY = pickle.load((, ))\nI would appreciate it if you could.
", "rawMarkdown": "pdnartreb \nThanks for sharing, and I really appreciate your decision.\nI have learned a lot from your notebook. If possible, can you public your dataset?\nnow, I cant reproduce your dataset\n\n```python\n%%time\nMETADATA = pickle.load(open(f'../final/METADATA.pkl', 'rb'))\ndataset = pickle.load(open(f'../data/processed/dataset.pkl', 'rb'))\ntokenizer_map = pickle.load(open(f'../data/processed/tokenizer_map.pkl', 'rb'))\npretrain_datasets = pickle.load(open(f'../data/processed/pretrain_datasets.pkl', 'rb'))\nY = pickle.load(open(f'../data/processed/Y.pkl', 'rb'))\n```\nI would appreciate it if you could.", "votes": null }, { "id": "2340161", "postDate": "07/11/2023 10:10:17", "content": "Glad you learned from our work @deepkun1995!
\nYou are indeed not able to reproduce because the code has been shared for information/educational and not for reproduction. We have not shared the data extraction and data building pipelines that consume a lot of resources.
\nI am not able to work anymore on the competition because I am moving out my house and have a lot to do. I will begin a world tour in 3 days. Sharing the documentation that we prepared for the host is possible but I would prefer to share it with the host first, if this competition is finalized one day… The host may share a 100% reproducable solution. If not, feel free to send an email via Kaggle and I will share a link to the documentation after the presentation to the host if they agree.
I will be in Japan in november and in Tokyo around 2023/11/15. I noticed that your team and you are from Japan. Do you know if there is a Kaggle community in Japan and Tokyo that would be interested in a meetup where I could present and share our solution for this competition? It would be a nice way for me to meet japonese people and maybe to be informed of nice things to do and to see in Japan.
", "rawMarkdown": "Glad you learned from our work @deepkun1995!\n\nYou are indeed not able to reproduce because the code has been shared for information/educational and not for reproduction. We have not shared the data extraction and data building pipelines that consume a lot of resources.\nI am not able to work anymore on the competition because I am moving out my house and have a lot to do. I will begin a world tour in 3 days. Sharing the documentation that we prepared for the host is possible but I would prefer to share it with the host first, if this competition is finalized one day... The host may share a 100% reproducable solution. If not, feel free to send an email via Kaggle and I will share a link to the documentation after the presentation to the host if they agree.\n\nI will be in Japan in november and in Tokyo around 2023/11/15. I noticed that your team and you are from Japan. Do you know if there is a Kaggle community in Japan and Tokyo that would be interested in a meetup where I could present and share our solution for this competition? It would be a nice way for me to meet japonese people and maybe to be informed of nice things to do and to see in Japan.", "votes": null }, { "id": "2342014", "postDate": "07/12/2023 12:43:01", "content": "Thanks for your reply.
\nI send you an email about meetup😃
\n@pdnartreb
I really appreciate your decision.
\nIt will be a very great source for people who want to dig into.
Amazing work, thanks for sharing this with us!
", "votes": null, "replies": [] }, { "id": 2325500, "author_name": "poojach7611", "author_url": "", "post_date": "07/01/2023 12:14:57", "content": "Thank you for sharing the code! It's great to see Kagglers contributing to the community by sharing their winning solutions. By making your code available, you not only give back to the community but also help in keeping machine learning an open field.
\nThe provided notebooks contain the code used to generate the final submission as well as the code for training and inference. It's worth noting that this code has not been executed on the Kaggle infrastructure and may require additional resources to run successfully.
\nOnce again, thank you for your contribution and for promoting openness in the field of machine learning!
", "votes": null, "replies": [] }, { "id": 2325657, "author_name": "akshayvyas02", "author_url": "", "post_date": "07/01/2023 14:01:07", "content": "@pdnartreb Thank you for sharing your code!Well done on your 1st place achievement!
", "votes": null, "replies": [] }, { "id": 2330061, "author_name": "andrestoledo92", "author_url": "", "post_date": "07/04/2023 17:29:32", "content": "Thanks for sharing learned a lot just reading through the code.
", "votes": null, "replies": [] }, { "id": 2335986, "author_name": "deepkun1995", "author_url": "", "post_date": "07/09/2023 02:47:03", "content": "@pdnartreb
\nThanks for sharing, and I really appreciate your decision.
\nI have learned a lot from your notebook. If possible, can you public your dataset?
\nnow, I cant reproduce your dataset
%%time\nMETADATA = pickle.load((, ))\ndataset = pickle.load((, ))\ntokenizer_map = pickle.load((, ))\npretrain_datasets = pickle.load((, ))\nY = pickle.load((, ))\nI would appreciate it if you could.
", "votes": null, "replies": [ { "id": 2340161, "author_name": "pdnartreb", "author_url": "", "post_date": "07/11/2023 10:10:17", "content": "Glad you learned from our work @deepkun1995!
\nYou are indeed not able to reproduce because the code has been shared for information/educational and not for reproduction. We have not shared the data extraction and data building pipelines that consume a lot of resources.
\nI am not able to work anymore on the competition because I am moving out my house and have a lot to do. I will begin a world tour in 3 days. Sharing the documentation that we prepared for the host is possible but I would prefer to share it with the host first, if this competition is finalized one day… The host may share a 100% reproducable solution. If not, feel free to send an email via Kaggle and I will share a link to the documentation after the presentation to the host if they agree.
I will be in Japan in november and in Tokyo around 2023/11/15. I noticed that your team and you are from Japan. Do you know if there is a Kaggle community in Japan and Tokyo that would be interested in a meetup where I could present and share our solution for this competition? It would be a nice way for me to meet japonese people and maybe to be informed of nice things to do and to see in Japan.
", "votes": null, "replies": [ { "id": 2342014, "author_name": "deepkun1995", "author_url": "", "post_date": "07/12/2023 12:43:01", "content": "Thanks for your reply.
\nI send you an email about meetup😃
\n@pdnartreb