Hi everyone!
\n\nI need some advise what I made wrong with making submission!?
\n\nMy simple approach with clear understanding :\n0. get all bird codes all_birds = df_train['ebird_code'].unique()\n1. read ../input/birdsong-recognition/test.csv\n2. iterate by rows with merging row_id and getting random value from all_birds
Here I have created notebook with demo. It has \" Failed\" :(
\n\nI need some advise from competitors or explanations from hosts, I think I have got right structure of submission file according to description.
\n\nP.S. We need use \"hidden\" test.csv, so loading prepared sample_submission.csv is not right way.
[UPDATE]\nThanks @cwthompson for providing kernel with sample submission using test.csv:
https://www.kaggle.com/cwthompson/birdsong-making-a-prediction
", "messages": [ { "id": "890721", "postDate": "06/17/2020 16:53:24", "content": "Hi everyone!
\n\nI need some advise what I made wrong with making submission!?
\n\nMy simple approach with clear understanding :\n0. get all bird codes all_birds = df_train['ebird_code'].unique()\n1. read ../input/birdsong-recognition/test.csv\n2. iterate by rows with merging row_id and getting random value from all_birds
Here I have created notebook with demo. It has \" Failed\" :(
\n\nI need some advise from competitors or explanations from hosts, I think I have got right structure of submission file according to description.
\n\nP.S. We need use \"hidden\" test.csv, so loading prepared sample_submission.csv is not right way.
[UPDATE]\nThanks @cwthompson for providing kernel with sample submission using test.csv:
https://www.kaggle.com/cwthompson/birdsong-making-a-prediction
", "rawMarkdown": "Hi everyone!\n\nI need some advise what I made wrong with making submission!?\n\nMy simple approach with clear understanding :\n0. get all bird codes `all_birds = df_train['ebird_code'].unique()`\n1. read `../input/birdsong-recognition/test.csv`\n2. iterate by rows with merging `row_id` and getting random value from `all_birds`\n\n\n[Here](https://www.kaggle.com/shonenkov/sample-submission?scriptVersionId=36605115) I have created notebook with demo. It has \" Failed\" :(\n\nI need some advise from competitors or explanations from hosts, I think I have got right structure of submission file according to description.\n\n\nP.S. We need use \"hidden\" test.csv, so loading prepared `sample_submission.csv` is not right way.\n\n[UPDATE]\nThanks @cwthompson for providing kernel with sample submission using `test.csv`:\n\nhttps://www.kaggle.com/cwthompson/birdsong-making-a-prediction", "votes": null }, { "id": "890754", "postDate": "06/17/2020 17:16:27", "content": "loading sample_submission.csv still works as it will get replaced by the file of appropriate lines. at least that's how I got my score. 0.0 for predicting all the classes btw
@carriesmi thank you for answer! but it doesnt help for making submission \"right way\"
", "rawMarkdown": "carriesmi thank you for answer! but it doesnt help for making submission \"right way\"", "votes": null }, { "id": "890809", "postDate": "06/17/2020 17:56:07", "content": "I ignored sample_submission.csv and created my own data frame. You need to use the row_ids from test.csv, and then another column called birds for your predictions.
\n\nIf this is not clear, I can make a public kernel to show you how to do it (just let me know).
", "rawMarkdown": "I ignored sample\\_submission.csv and created my own data frame. You need to use the row_ids from test.csv, and then another column called birds for your predictions.\n\nIf this is not clear, I can make a public kernel to show you how to do it (just let me know).", "votes": null }, { "id": "890814", "postDate": "06/17/2020 17:58:30", "content": "yes, I also have created own dataframe:
\n\nhttps://www.kaggle.com/shonenkov/sample-submission?scriptVersionId=36605115
\n\n@cwthompson Could you make kernel with your sample submission? I think It will be very usefull for all 😃
\n\nThank you!
", "rawMarkdown": "yes, I also have created own dataframe:\n\nhttps://www.kaggle.com/shonenkov/sample-submission?scriptVersionId=36605115\n\n@cwthompson Could you make kernel with your sample submission? I think It will be very usefull for all 😃 \n\nThank you!", "votes": null }, { "id": "890820", "postDate": "06/17/2020 18:02:48", "content": "I've made it public here:\nhttps://www.kaggle.com/cwthompson/birdsong-making-a-prediction\nPlease also note my comment at the bottom if you are trying to submit.
\n\nI will have a look at yours in a min to see what went wrong.
", "rawMarkdown": "I've made it public here:\nhttps://www.kaggle.com/cwthompson/birdsong-making-a-prediction\nPlease also note my comment at the bottom if you are trying to submit.\n\nI will have a look at yours in a min to see what went wrong.", "votes": null }, { "id": "890833", "postDate": "06/17/2020 18:08:58", "content": "thank you @cwthompson. really helped!
", "rawMarkdown": "thank you @cwthompson. really helped!", "votes": null }, { "id": "995767", "postDate": "09/02/2020 18:46:53", "content": "in your code you have:
\n\n\nHIDDEN_TEST = len(glob('../input/test/*.mp4')) != 0
\n
Shouldn't it be:
\nHIDDEN_TEST = len(glob('../input/test/*.mp3')) != 0
", "rawMarkdown": "in your code you have:\n\n> HIDDEN_TEST = len(glob('../input/test/*.mp4')) != 0\n\nShouldn't it be:\n\nHIDDEN_TEST = len(glob('../input/test/*.mp3')) != 0", "votes": null }, { "id": "995769", "postDate": "09/02/2020 18:47:52", "content": "Predicting all the classes should not result in a score of 0.
", "rawMarkdown": "Predicting all the classes should not result in a score of 0.", "votes": null } ], "comments": [ { "id": 995767, "author_name": "cpmpml", "author_url": "", "post_date": "09/02/2020 18:46:53", "content": "in your code you have:
\n\n\nHIDDEN_TEST = len(glob('../input/test/*.mp4')) != 0
\n
Shouldn't it be:
\nHIDDEN_TEST = len(glob('../input/test/*.mp3')) != 0
", "votes": null, "replies": [] }, { "id": 890754, "author_name": "carriesmi", "author_url": "", "post_date": "06/17/2020 17:16:27", "content": "loading sample_submission.csv still works as it will get replaced by the file of appropriate lines. at least that's how I got my score. 0.0 for predicting all the classes btw
@carriesmi thank you for answer! but it doesnt help for making submission \"right way\"
", "votes": null, "replies": [] }, { "id": 995769, "author_name": "cpmpml", "author_url": "", "post_date": "09/02/2020 18:47:52", "content": "Predicting all the classes should not result in a score of 0.
", "votes": null, "replies": [] } ] }, { "id": 890809, "author_name": "cwthompson", "author_url": "", "post_date": "06/17/2020 17:56:07", "content": "I ignored sample_submission.csv and created my own data frame. You need to use the row_ids from test.csv, and then another column called birds for your predictions.
\n\nIf this is not clear, I can make a public kernel to show you how to do it (just let me know).
", "votes": null, "replies": [ { "id": 890814, "author_name": "shonenkov", "author_url": "", "post_date": "06/17/2020 17:58:30", "content": "yes, I also have created own dataframe:
\n\nhttps://www.kaggle.com/shonenkov/sample-submission?scriptVersionId=36605115
\n\n@cwthompson Could you make kernel with your sample submission? I think It will be very usefull for all 😃
\n\nThank you!
", "votes": null, "replies": [] }, { "id": 890820, "author_name": "cwthompson", "author_url": "", "post_date": "06/17/2020 18:02:48", "content": "I've made it public here:\nhttps://www.kaggle.com/cwthompson/birdsong-making-a-prediction\nPlease also note my comment at the bottom if you are trying to submit.
\n\nI will have a look at yours in a min to see what went wrong.
", "votes": null, "replies": [] }, { "id": 890833, "author_name": "shonenkov", "author_url": "", "post_date": "06/17/2020 18:08:58", "content": "thank you @cwthompson. really helped!
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "890721": "Hi everyone!\n\nI need some advise what I made wrong with making submission!?\n\nMy simple approach with clear understanding :\n0. get all bird codes `all_birds = df_train['ebird_code'].unique()`\n1. read `../input/birdsong-recognition/test.csv`\n2. iterate by rows with merging `row_id` and getting random value from `all_birds`\n\n\n[Here](https://www.kaggle.com/shonenkov/sample-submission?scriptVersionId=36605115) I have created notebook with demo. It has \" Failed\" :(\n\nI need some advise from competitors or explanations from hosts, I think I have got right structure of submission file according to description.\n\n\nP.S. We need use \"hidden\" test.csv, so loading prepared `sample_submission.csv` is not right way.\n\n[UPDATE]\nThanks @cwthompson for providing kernel with sample submission using `test.csv`:\n\nhttps://www.kaggle.com/cwthompson/birdsong-making-a-prediction", "890754": "loading `sample_submission.csv` still works as it will get replaced by the file of appropriate lines. at least that's how I got my score. 0.0 for predicting all the classes btw", "890803": "carriesmi thank you for answer! but it doesnt help for making submission \"right way\"", "890809": "I ignored sample\\_submission.csv and created my own data frame. You need to use the row_ids from test.csv, and then another column called birds for your predictions.\n\nIf this is not clear, I can make a public kernel to show you how to do it (just let me know).", "890814": "yes, I also have created own dataframe:\n\nhttps://www.kaggle.com/shonenkov/sample-submission?scriptVersionId=36605115\n\n@cwthompson Could you make kernel with your sample submission? I think It will be very usefull for all 😃 \n\nThank you!", "890820": "I've made it public here:\nhttps://www.kaggle.com/cwthompson/birdsong-making-a-prediction\nPlease also note my comment at the bottom if you are trying to submit.\n\nI will have a look at yours in a min to see what went wrong.", "890833": "thank you @cwthompson. really helped!", "995767": "in your code you have:\n\n> HIDDEN_TEST = len(glob('../input/test/*.mp4')) != 0\n\nShouldn't it be:\n\nHIDDEN_TEST = len(glob('../input/test/*.mp3')) != 0", "995769": "Predicting all the classes should not result in a score of 0." }, "source": "meta" }