{ "id": 369706, "title": "3 resources to get started with Computer Vision in this competition 🚀🚀🚀", "url": "/competitions/rsna-breast-cancer-detection/discussion/369706", "author_name": "Radek Osmulski", "post_date": "2022-12-01T06:07:29.006000", "votes": 26, "comment_count": 0, "views": 0, "content": "

Hey!

\n

I came across questions on the forums on how to get started with computer vision (and specifically this competition) and wanted to share a couple of resources.

\n

1. The absolutely best course on practical computer vision

\n

The absolutely best course for CV is Practical Deep Learning for Coders (2022) by fast.ai. The first 3 lectures should give you enough background to become dangerous in this competition.

\n

I won a Kaggle competition straight after taking an earlier version of this course. You can read more about this here: [1st place solution] pretrained CNNs -> xgboost -> F1 optimization

\n

2. A course with a more academic twist

\n

If you prefer to learn in a more traditional fashion, Deep Learning - Stanford CS231N is your best bet! A course by Stanford, brought to us by the people behind Imagenet! Highly recommended.

\n

3. An interactive course with exercises

\n

If you prefer lectures broken into smaller, self-contained videos and would like to have access to homework/practice quizzes to check and solidify your understanding, the Machine Learning Specialization by Andrew NG is a fantastic choice!

\n

DO NOTE: If you are hoping to learn enough to become dangerous in this competition, PICK JUST ONE COURSE. I cannot stress this enough.

\n

In fact, I ordered the courses in the order of how practical they are. To get up and running quickly in this competition the practical aspects of training CV models are of utter importance. You can acquire more theory as you are waiting for your models to train 😇 (that is, once you get them training! and for this you need a good introduction!

\n

Happy Kaggling! 🥳

\n

Other resources you might find useful:

\n", "messages": [ { "id": 2050976, "postDate": "2022-12-01T06:07:29.007Z", "content": "

Hey!

\n

I came across questions on the forums on how to get started with computer vision (and specifically this competition) and wanted to share a couple of resources.

\n

1. The absolutely best course on practical computer vision

\n

The absolutely best course for CV is Practical Deep Learning for Coders (2022) by fast.ai. The first 3 lectures should give you enough background to become dangerous in this competition.

\n

I won a Kaggle competition straight after taking an earlier version of this course. You can read more about this here: [1st place solution] pretrained CNNs -> xgboost -> F1 optimization

\n

2. A course with a more academic twist

\n

If you prefer to learn in a more traditional fashion, Deep Learning - Stanford CS231N is your best bet! A course by Stanford, brought to us by the people behind Imagenet! Highly recommended.

\n

3. An interactive course with exercises

\n

If you prefer lectures broken into smaller, self-contained videos and would like to have access to homework/practice quizzes to check and solidify your understanding, the Machine Learning Specialization by Andrew NG is a fantastic choice!

\n

DO NOTE: If you are hoping to learn enough to become dangerous in this competition, PICK JUST ONE COURSE. I cannot stress this enough.

\n

In fact, I ordered the courses in the order of how practical they are. To get up and running quickly in this competition the practical aspects of training CV models are of utter importance. You can acquire more theory as you are waiting for your models to train 😇 (that is, once you get them training! and for this you need a good introduction!

\n

Happy Kaggling! 🥳

\n

Other resources you might find useful:

\n", "rawMarkdown": "Hey!\n\nI came across questions on the forums on how to get started with computer vision (and specifically this competition) and wanted to share a couple of resources.\n\n# 1. The absolutely best course on *practical* computer vision\n\nThe absolutely best course for CV is [Practical Deep Learning for Coders (2022)](https://course.fast.ai/) by [fast.ai](https://fast.ai). The first 3 lectures should give you enough background to become dangerous in this competition.\n\nI won a Kaggle competition straight after taking an earlier version of this course. You can read more about this here: [[1st place solution] pretrained CNNs -> xgboost -> F1 optimization](https://www.kaggle.com/competitions/imaterialist-challenge-fashion-2018/discussion/57944)\n\n# 2. A course with a more academic twist\n\nIf you prefer to learn in a more traditional fashion, [Deep Learning - Stanford CS231N](https://www.youtube.com/playlist?list=PLSVEhWrZWDHQTBmWZufjxpw3s8sveJtnJ) is your best bet! A course by Stanford, brought to us by the people behind Imagenet! Highly recommended.\n\n# 3. An interactive course with exercises\n\nIf you prefer lectures broken into smaller, self-contained videos and would like to have access to homework/practice quizzes to check and solidify your understanding, the [Machine Learning Specialization](https://www.coursera.org/specializations/machine-learning-introduction#courses) by Andrew NG is a fantastic choice!\n\n**DO NOTE:** If you are hoping to learn enough to become dangerous in this competition, PICK JUST ONE COURSE. I cannot stress this enough.\n\nIn fact, I ordered the courses in the order of how practical they are. To get up and running quickly in this competition the practical aspects of training CV models are of utter importance. You can acquire more theory as you are waiting for your models to train 😇 (that is, once you get them training! and for this you need a good introduction!\n\nHappy Kaggling! 🥳\n\n### Other resources you might find useful:\n\n* [💡 how to process DICOM images to PNGs](https://www.kaggle.com/code/radek1/how-to-process-dicom-images-to-pngs)\n* [📊 EDA + training a fast.ai model + submission 🚀](https://www.kaggle.com/code/radek1/eda-training-a-fast-ai-model-submission)\n* [🤖 [fast.ai starter pack] train + inference 🚀](https://www.kaggle.com/code/radek1/fast-ai-starter-pack-train-inference)\n* [📸 Over 56GB of processed data, 5 different methods 🥳](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/371268)\n* [3 resources to get started with Computer Vision in this competition 🚀🚀🚀](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/369706)\n* [💡 6 Computer Vision tricks for faster training and better models 🚀🚀🚀](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/369155)\n", "votes": 26 } ], "comments": [], "raw_markdown_by_id": { "2050976": "Hey!\n\nI came across questions on the forums on how to get started with computer vision (and specifically this competition) and wanted to share a couple of resources.\n\n# 1. The absolutely best course on *practical* computer vision\n\nThe absolutely best course for CV is [Practical Deep Learning for Coders (2022)](https://course.fast.ai/) by [fast.ai](https://fast.ai). The first 3 lectures should give you enough background to become dangerous in this competition.\n\nI won a Kaggle competition straight after taking an earlier version of this course. You can read more about this here: [[1st place solution] pretrained CNNs -> xgboost -> F1 optimization](https://www.kaggle.com/competitions/imaterialist-challenge-fashion-2018/discussion/57944)\n\n# 2. A course with a more academic twist\n\nIf you prefer to learn in a more traditional fashion, [Deep Learning - Stanford CS231N](https://www.youtube.com/playlist?list=PLSVEhWrZWDHQTBmWZufjxpw3s8sveJtnJ) is your best bet! A course by Stanford, brought to us by the people behind Imagenet! Highly recommended.\n\n# 3. An interactive course with exercises\n\nIf you prefer lectures broken into smaller, self-contained videos and would like to have access to homework/practice quizzes to check and solidify your understanding, the [Machine Learning Specialization](https://www.coursera.org/specializations/machine-learning-introduction#courses) by Andrew NG is a fantastic choice!\n\n**DO NOTE:** If you are hoping to learn enough to become dangerous in this competition, PICK JUST ONE COURSE. I cannot stress this enough.\n\nIn fact, I ordered the courses in the order of how practical they are. To get up and running quickly in this competition the practical aspects of training CV models are of utter importance. You can acquire more theory as you are waiting for your models to train 😇 (that is, once you get them training! and for this you need a good introduction!\n\nHappy Kaggling! 🥳\n\n### Other resources you might find useful:\n\n* [💡 how to process DICOM images to PNGs](https://www.kaggle.com/code/radek1/how-to-process-dicom-images-to-pngs)\n* [📊 EDA + training a fast.ai model + submission 🚀](https://www.kaggle.com/code/radek1/eda-training-a-fast-ai-model-submission)\n* [🤖 [fast.ai starter pack] train + inference 🚀](https://www.kaggle.com/code/radek1/fast-ai-starter-pack-train-inference)\n* [📸 Over 56GB of processed data, 5 different methods 🥳](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/371268)\n* [3 resources to get started with Computer Vision in this competition 🚀🚀🚀](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/369706)\n* [💡 6 Computer Vision tricks for faster training and better models 🚀🚀🚀](https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/369155)\n" } }