{ "id": 206137, "title": "Speeding Up-Preprocessing 40%: Lycon vs. Albumentations", "url": "/competitions/ranzcr-clip-catheter-line-classification/discussion/206137", "author_name": "Zac Dannelly", "post_date": "2020-12-23T11:42:12.661000", "votes": 5, "comment_count": 6, "views": 0, "content": "

I have traditionally seen albumentations as one of the most common pre-processing libraries in these sorts of competitoions. Howver, recently I have also come across lycon which at least for initial re-sizeing seems to have a speed up advantage.

", "messages": [ { "id": 1123634, "postDate": "2020-12-23T11:42:12.660Z", "content": "

I have traditionally seen albumentations as one of the most common pre-processing libraries in these sorts of competitoions. Howver, recently I have also come across lycon which at least for initial re-sizeing seems to have a speed up advantage.

", "rawMarkdown": "I have traditionally seen [albumentations](https://github.com/albumentations-team/albumentations) as one of the most common pre-processing libraries in these sorts of competitoions. Howver, recently I have also come across [lycon](https://github.com/ethereon/lycon) which at least for initial re-sizeing seems to have a speed up advantage.", "votes": 5 }, { "id": 1123911, "postDate": "2020-12-23T15:22:13.460Z", "content": "

Thanks for sharing! According to their benchmarks its pretty much even with opencv so albumentations is probably as fast and gives more options

", "rawMarkdown": "Thanks for sharing! According to their benchmarks its pretty much even with opencv so albumentations is probably as fast and gives more options ", "votes": 3 }, { "id": 1123940, "postDate": "2020-12-23T15:47:33.253Z", "content": "

Is there a dataset in Kaggle with that library?

", "rawMarkdown": "Is there a dataset in Kaggle with that library?", "replies": [ { "id": 1124411, "postDate": "2020-12-23T23:05:42.617Z", "content": "

Gave a quick search but didn't see any. I will give a look across some datasets and maybe throw a notebook together over the weekend.

", "rawMarkdown": "Gave a quick search but didn't see any. I will give a look across some datasets and maybe throw a notebook together over the weekend." }, { "id": 1144775, "postDate": "2021-01-08T17:01:52.497Z", "content": "

Lycon was used pretty extensively by a number of people in the steganography competition, so check for kernels there,

", "rawMarkdown": "Lycon was used pretty extensively by a number of people in the steganography competition, so check for kernels there,", "votes": 1 } ] }, { "id": 1124410, "postDate": "2020-12-23T23:05:33.293Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1124933, "postDate": "2020-12-24T09:27:21.297Z", "content": "

Thanks for the information @dannellyz

", "rawMarkdown": "Thanks for the information @dannellyz ", "votes": 1 } ], "comments": [ { "id": 1123911, "author_name": "Yann Majewski", "author_url": "", "post_date": "2020-12-23T15:22:13.460000", "content": "

Thanks for sharing! According to their benchmarks its pretty much even with opencv so albumentations is probably as fast and gives more options

", "votes": 3, "replies": [] }, { "id": 1123940, "author_name": "Hiram Coria 🧬", "author_url": "", "post_date": "2020-12-23T15:47:33.253000", "content": "

Is there a dataset in Kaggle with that library?

", "votes": 0, "replies": [ { "id": 1124411, "author_name": "Zac Dannelly", "author_url": "", "post_date": "2020-12-23T23:05:42.617000", "content": "

Gave a quick search but didn't see any. I will give a look across some datasets and maybe throw a notebook together over the weekend.

", "votes": 0, "replies": [] }, { "id": 1144775, "author_name": "عثمان", "author_url": "", "post_date": "2021-01-08T17:01:52.497000", "content": "

Lycon was used pretty extensively by a number of people in the steganography competition, so check for kernels there,

", "votes": 1, "replies": [] } ] }, { "id": 1124410, "author_name": "", "author_url": "", "post_date": "2020-12-23T23:05:33.293000", "content": "", "votes": 0, "replies": [] }, { "id": 1124933, "author_name": "Saurabh Shahane", "author_url": "", "post_date": "2020-12-24T09:27:21.297000", "content": "

Thanks for the information @dannellyz

", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "1123634": "I have traditionally seen [albumentations](https://github.com/albumentations-team/albumentations) as one of the most common pre-processing libraries in these sorts of competitoions. Howver, recently I have also come across [lycon](https://github.com/ethereon/lycon) which at least for initial re-sizeing seems to have a speed up advantage.", "1123911": "Thanks for sharing! According to their benchmarks its pretty much even with opencv so albumentations is probably as fast and gives more options ", "1123940": "Is there a dataset in Kaggle with that library?", "1124410": "", "1124933": "Thanks for the information @dannellyz " } }