{ "id": 202485, "title": "How To: Reuse Previous Competition Code On New Competition", "url": "/competitions/cassava-leaf-disease-classification/discussion/202485", "author_name": "", "post_date": "2020-12-10T09:20:35.969312200Z", "votes": 24, "comment_count": 3, "views": 0, "content": "

Hi all,

\n

I've publish a baseline notebook here just now:

\n

https://www.kaggle.com/haqishen/baseline-modified-from-previous-competition

\n

But the key word here is not baseline , it's Reuse

\n

I copy my notebook shared on previous competition (PANDA) then modified a little bit to build this baseline.

\n

That only cost me around 20min.

\n

If you're interested in how to reuse code please have a check ;)

", "messages": [ { "id": "1108109", "postDate": "12/10/2020 09:20:35", "content": "

Hi all,

\n

I've publish a baseline notebook here just now:

\n

https://www.kaggle.com/haqishen/baseline-modified-from-previous-competition

\n

But the key word here is not baseline , it's Reuse

\n

I copy my notebook shared on previous competition (PANDA) then modified a little bit to build this baseline.

\n

That only cost me around 20min.

\n

If you're interested in how to reuse code please have a check ;)

", "rawMarkdown": "Hi all,\n\nI've publish a baseline notebook here just now:\n\nhttps://www.kaggle.com/haqishen/baseline-modified-from-previous-competition\n\nBut the key word here is not *baseline* , it's **Reuse**\n\nI copy [my notebook shared on previous competition](https://www.kaggle.com/haqishen/train-efficientnet-b0-w-36-tiles-256-lb0-87) (PANDA) then modified a little bit to build this baseline.\n\nThat only cost me around 20min.\n\nIf you're interested in how to reuse code please have a check ;)", "votes": null }, { "id": "1108167", "postDate": "12/10/2020 10:56:43", "content": "

Thanks for sharing a nice notebook.
\nI learned a lot from the tips and codes you(and your team) shared in other competitions.

\n

@haqishen
\np.s. I want to release my work sooner or later, but I still need time.

", "rawMarkdown": "Thanks for sharing a nice notebook.\nI learned a lot from the tips and codes you(and your team) shared in other competitions.\n\n@haqishen \np.s. I want to release my work sooner or later, but I still need time.", "votes": null }, { "id": "1109041", "postDate": "12/11/2020 09:18:15", "content": "

I find the old dataset(2019) is a crop of new dataset(2020). what about take it as a bounding box to train a object detection model for the dataset 2020

", "rawMarkdown": "I find the old dataset(2019) is a crop of new dataset(2020). what about take it as a bounding box to train a object detection model for the dataset 2020", "votes": null }, { "id": "1109929", "postDate": "12/12/2020 08:12:59", "content": "

Thanks @haqishen for sharing your work. BTW, have you considered augmenting the previous year's same competition data to this one?? Might be interesting to see that!

", "rawMarkdown": "Thanks @haqishen for sharing your work. BTW, have you considered augmenting the previous year's same competition data to this one?? Might be interesting to see that!", "votes": null } ], "comments": [ { "id": 1108167, "author_name": "piantic", "author_url": "", "post_date": "12/10/2020 10:56:43", "content": "

Thanks for sharing a nice notebook.
\nI learned a lot from the tips and codes you(and your team) shared in other competitions.

\n

@haqishen
\np.s. I want to release my work sooner or later, but I still need time.

", "votes": null, "replies": [] }, { "id": 1109041, "author_name": "raininbox", "author_url": "", "post_date": "12/11/2020 09:18:15", "content": "

I find the old dataset(2019) is a crop of new dataset(2020). what about take it as a bounding box to train a object detection model for the dataset 2020

", "votes": null, "replies": [] }, { "id": 1109929, "author_name": "divyansh22", "author_url": "", "post_date": "12/12/2020 08:12:59", "content": "

Thanks @haqishen for sharing your work. BTW, have you considered augmenting the previous year's same competition data to this one?? Might be interesting to see that!

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1108109": "Hi all,\n\nI've publish a baseline notebook here just now:\n\nhttps://www.kaggle.com/haqishen/baseline-modified-from-previous-competition\n\nBut the key word here is not *baseline* , it's **Reuse**\n\nI copy [my notebook shared on previous competition](https://www.kaggle.com/haqishen/train-efficientnet-b0-w-36-tiles-256-lb0-87) (PANDA) then modified a little bit to build this baseline.\n\nThat only cost me around 20min.\n\nIf you're interested in how to reuse code please have a check ;)", "1108167": "Thanks for sharing a nice notebook.\nI learned a lot from the tips and codes you(and your team) shared in other competitions.\n\n@haqishen \np.s. I want to release my work sooner or later, but I still need time.", "1109041": "I find the old dataset(2019) is a crop of new dataset(2020). what about take it as a bounding box to train a object detection model for the dataset 2020", "1109929": "Thanks @haqishen for sharing your work. BTW, have you considered augmenting the previous year's same competition data to this one?? Might be interesting to see that!" }, "source": "meta" }