\n
![\"\"](\"https://i.ibb.co/fqQshCc/0-444.png\")
I stopped working on the competition a few weeks ago and I forgot that I have to choose the notebooks for private LB! I want to start by saying thank you Coleridge Initiative for hosting such an interesting competition and thank you to the community that makes Kaggle such a friendly, cozy place to learn data science.
\nThis is my first competition on Kaggle and my first experience in NLP. I came into the competition not knowing what to expect. I came out with a bronze medal but most importantly, the knowledge I learned from so many people. Special thanks to @tungmphung for sharing great notebooks that I learned so much from at the beginning.
\nAs the title says, I could've got a gold medal. I had a notebook with private LB 0.444
\n I stopped working on the competition a few weeks ago and I forgot that I have to choose the notebooks for private LB!
In the end, the bronze medal I got was from a LB probing notebook 😂 lol
\nwhat I did in the would've-been-13th-place-notebook was really simple:
\nI was having mixed feelings about losing a gold medal for a few moments, but I had a great time working on this competition and learning from every notebook and discussion, what more can I ask for?
", "messages": [ { "id": 1362723, "postDate": "2021-06-23T15:30:32.227Z", "content": "I want to start by saying thank you Coleridge Initiative for hosting such an interesting competition and thank you to the community that makes Kaggle such a friendly, cozy place to learn data science.
\nThis is my first competition on Kaggle and my first experience in NLP. I came into the competition not knowing what to expect. I came out with a bronze medal but most importantly, the knowledge I learned from so many people. Special thanks to @tungmphung for sharing great notebooks that I learned so much from at the beginning.
\nAs the title says, I could've got a gold medal. I had a notebook with private LB 0.444
\n I stopped working on the competition a few weeks ago and I forgot that I have to choose the notebooks for private LB!
In the end, the bronze medal I got was from a LB probing notebook 😂 lol
\nwhat I did in the would've-been-13th-place-notebook was really simple:
\nI was having mixed feelings about losing a gold medal for a few moments, but I had a great time working on this competition and learning from every notebook and discussion, what more can I ask for?
", "rawMarkdown": "I want to start by saying thank you Coleridge Initiative for hosting such an interesting competition and thank you to the community that makes Kaggle such a friendly, cozy place to learn data science. \n\nThis is my first competition on Kaggle and my first experience in NLP. I came into the competition not knowing what to expect. I came out with a bronze medal but most importantly, the knowledge I learned from so many people. Special thanks to @tungmphung for sharing great notebooks that I learned so much from at the beginning.\n\nAs the title says, I could've got a gold medal. I had a notebook with private LB 0.444\n I stopped working on the competition a few weeks ago and I forgot that I have to choose the notebooks for private LB! \n\nIn the end, the bronze medal I got was from a LB probing notebook 😂 lol\n\nwhat I did in the would've-been-13th-place-notebook was really simple:\n1. divide train data into 256 tokens long chunks\n2. use all chunks that have positive labels in them and discard half of the chunks with no labels\n3. take XLM-Roberta finetuned on conll03, from dense layer copy weights for only \"O\" (out of entity) and \"MISC\" label and use \"MISC\" for positive in binary classification head\n4. train 3 epoch\n\nI was having mixed feelings about losing a gold medal for a few moments, but I had a great time working on this competition and learning from every notebook and discussion, what more can I ask for?", "votes": 14 }, { "id": 1366004, "postDate": "2021-06-26T11:14:40.600Z", "content": "can you please tell me where did you learn nlp from . Because i am a beginner and i want a dig deep into this field and hopefully solve some kaggle problems like you
", "rawMarkdown": "can you please tell me where did you learn nlp from . Because i am a beginner and i want a dig deep into this field and hopefully solve some kaggle problems like you", "votes": 1, "replies": [ { "id": 1366375, "postDate": "2021-06-26T17:47:25.747Z", "content": "I am still very much a beginner myself, so I don't know if I could help you much. But if you want my 2 cents, here it is.
\nPublic kernels are very good for beginners. I went through all the popular ones, picked out ones that made more sense to me and tried to replicate them.
\nHuggingface source code is a great place to learn when implementing a training loop and adding custom features.
\nTowards data science and medium are good sources to learn as well. Posts generally consist of easy-to-understand explanations and illustrations; check them out first when it shows on your google search.
\nResearch papers, this is probably the most important one.
", "rawMarkdown": "I am still very much a beginner myself, so I don't know if I could help you much. But if you want my 2 cents, here it is. \n\nPublic kernels are very good for beginners. I went through all the popular ones, picked out ones that made more sense to me and tried to replicate them. \n\nHuggingface source code is a great place to learn when implementing a training loop and adding custom features. \n\nTowards data science and medium are good sources to learn as well. Posts generally consist of easy-to-understand explanations and illustrations; check them out first when it shows on your google search.\n\nResearch papers, this is probably the most important one. ", "votes": 1 } ] }, { "id": 1367856, "postDate": "2021-06-28T06:07:14.650Z", "content": "A pity u missed that well-deserved gold! Anyway congrats and BTW there is a similar comp - CommonLit which you may want to try out. There is still a month left for that…
", "rawMarkdown": "A pity u missed that well-deserved gold! Anyway congrats and BTW there is a similar comp - CommonLit which you may want to try out. There is still a month left for that...", "votes": 2, "replies": [ { "id": 1396383, "postDate": "2021-07-22T04:30:24.400Z", "content": "Looks like you took my advice :) Happy to see u on top of CommonLit
", "rawMarkdown": "Looks like you took my advice :) Happy to see u on top of CommonLit" }, { "id": 1396419, "postDate": "2021-07-22T05:23:46.547Z", "content": "Thank you for your kind words. I hope the shake up won’t be as horrible as coleridge😂
", "rawMarkdown": "Thank you for your kind words. I hope the shake up won’t be as horrible as coleridge😂", "votes": 1 } ] }, { "id": 1398326, "postDate": "2021-07-24T03:55:05.797Z", "content": "Don't worry buddy. many many golds coming soon.
", "rawMarkdown": "Don't worry buddy. many many golds coming soon." }, { "id": 1366675, "postDate": "2021-06-27T05:01:26.083Z", "content": "It is the journey that matters at kaggle @potatoc5 ! Medals are just the destination, as long as you are enjoying the process, it is fine. All the best!
", "rawMarkdown": "It is the journey that matters at kaggle @potatoc5 ! Medals are just the destination, as long as you are enjoying the process, it is fine. All the best!" }, { "id": 1362727, "postDate": "2021-06-23T15:32:36.650Z", "rawMarkdown": "", "votes": -1, "isDeleted": true, "replies": [ { "id": 1362863, "postDate": "2021-06-23T17:40:50.370Z", "content": "In my opinion, the goal is not really to get the best result, but to build a reliable model that will perform well on unseen data. Selecting your submissions should not be a guessing game :)
", "rawMarkdown": "In my opinion, the goal is not really to get the best result, but to build a reliable model that will perform well on unseen data. Selecting your submissions should not be a guessing game :)", "votes": 3 }, { "id": 1362869, "postDate": "2021-06-23T17:51:58.697Z", "rawMarkdown": "", "votes": -1, "isDeleted": true } ] } ], "comments": [ { "id": 1366004, "author_name": "killua", "author_url": "", "post_date": "2021-06-26T11:14:40.600000", "content": "can you please tell me where did you learn nlp from . Because i am a beginner and i want a dig deep into this field and hopefully solve some kaggle problems like you
", "votes": 1, "replies": [ { "id": 1366375, "author_name": "hxcc", "author_url": "", "post_date": "2021-06-26T17:47:25.747000", "content": "I am still very much a beginner myself, so I don't know if I could help you much. But if you want my 2 cents, here it is.
\nPublic kernels are very good for beginners. I went through all the popular ones, picked out ones that made more sense to me and tried to replicate them.
\nHuggingface source code is a great place to learn when implementing a training loop and adding custom features.
\nTowards data science and medium are good sources to learn as well. Posts generally consist of easy-to-understand explanations and illustrations; check them out first when it shows on your google search.
\nResearch papers, this is probably the most important one.
", "votes": 1, "replies": [] } ] }, { "id": 1367856, "author_name": "Allohvk", "author_url": "", "post_date": "2021-06-28T06:07:14.650000", "content": "A pity u missed that well-deserved gold! Anyway congrats and BTW there is a similar comp - CommonLit which you may want to try out. There is still a month left for that…
", "votes": 2, "replies": [ { "id": 1396383, "author_name": "Allohvk", "author_url": "", "post_date": "2021-07-22T04:30:24.400000", "content": "Looks like you took my advice :) Happy to see u on top of CommonLit
", "votes": 0, "replies": [] }, { "id": 1396419, "author_name": "hxcc", "author_url": "", "post_date": "2021-07-22T05:23:46.547000", "content": "Thank you for your kind words. I hope the shake up won’t be as horrible as coleridge😂
", "votes": 1, "replies": [] } ] }, { "id": 1398326, "author_name": "Vikram Sandu", "author_url": "", "post_date": "2021-07-24T03:55:05.797000", "content": "Don't worry buddy. many many golds coming soon.
", "votes": 0, "replies": [] }, { "id": 1366675, "author_name": "Old Monk", "author_url": "", "post_date": "2021-06-27T05:01:26.083000", "content": "It is the journey that matters at kaggle @potatoc5 ! Medals are just the destination, as long as you are enjoying the process, it is fine. All the best!
", "votes": 0, "replies": [] }, { "id": 1362727, "author_name": "", "author_url": "", "post_date": "2021-06-23T15:32:36.650000", "content": "", "votes": -1, "replies": [ { "id": 1362863, "author_name": "Theo Viel", "author_url": "", "post_date": "2021-06-23T17:40:50.370000", "content": "In my opinion, the goal is not really to get the best result, but to build a reliable model that will perform well on unseen data. Selecting your submissions should not be a guessing game :)
", "votes": 3, "replies": [] }, { "id": 1362869, "author_name": "", "author_url": "", "post_date": "2021-06-23T17:51:58.697000", "content": "", "votes": -1, "replies": [] } ] } ], "raw_markdown_by_id": { "1362723": "I want to start by saying thank you Coleridge Initiative for hosting such an interesting competition and thank you to the community that makes Kaggle such a friendly, cozy place to learn data science. \n\nThis is my first competition on Kaggle and my first experience in NLP. I came into the competition not knowing what to expect. I came out with a bronze medal but most importantly, the knowledge I learned from so many people. Special thanks to @tungmphung for sharing great notebooks that I learned so much from at the beginning.\n\nAs the title says, I could've got a gold medal. I had a notebook with private LB 0.444\n I stopped working on the competition a few weeks ago and I forgot that I have to choose the notebooks for private LB! \n\nIn the end, the bronze medal I got was from a LB probing notebook 😂 lol\n\nwhat I did in the would've-been-13th-place-notebook was really simple:\n1. divide train data into 256 tokens long chunks\n2. use all chunks that have positive labels in them and discard half of the chunks with no labels\n3. take XLM-Roberta finetuned on conll03, from dense layer copy weights for only \"O\" (out of entity) and \"MISC\" label and use \"MISC\" for positive in binary classification head\n4. train 3 epoch\n\nI was having mixed feelings about losing a gold medal for a few moments, but I had a great time working on this competition and learning from every notebook and discussion, what more can I ask for?", "1366004": "can you please tell me where did you learn nlp from . Because i am a beginner and i want a dig deep into this field and hopefully solve some kaggle problems like you", "1367856": "A pity u missed that well-deserved gold! Anyway congrats and BTW there is a similar comp - CommonLit which you may want to try out. There is still a month left for that...", "1398326": "Don't worry buddy. many many golds coming soon.", "1366675": "It is the journey that matters at kaggle @potatoc5 ! Medals are just the destination, as long as you are enjoying the process, it is fine. All the best!", "1362727": "" } }