์ฝ˜ํ…์ธ ๋กœ ๊ฑด๋„ˆ๋›ฐ๊ธฐ

A Guide on Using Kaggle to Train Your YOLO11 Models

If you are learning about AI and working on small projects, you might not have access to powerful computing resources yet, and high-end hardware can be pretty expensive. Fortunately, Kaggle, a platform owned by Google, offers a great solution. Kaggle provides a free, cloud-based environment where you can access GPU resources, handle large datasets, and collaborate with a diverse community of data scientists and machine learning enthusiasts.

Kaggle is a great choice for training and experimenting with Ultralytics YOLO11 models. Kaggle Notebooks make using popular machine-learning libraries and frameworks in your projects easy. Let's explore Kaggle's main features and learn how you can train YOLO11 models on this platform!

Kaggle์ด๋ž€ ๋ฌด์—‡์ธ๊ฐ€์š”?

Kaggle์€ ์ „ ์„ธ๊ณ„์˜ ๋ฐ์ดํ„ฐ ๊ณผํ•™์ž๋“ค์ด ๋ชจ์—ฌ ์‹ค์ œ ๋ฐ์ดํ„ฐ ๊ณผํ•™ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๊ธฐ ์œ„ํ•ด ํ˜‘์—…ํ•˜๊ณ , ๋ฐฐ์šฐ๊ณ , ๊ฒฝ์Ÿํ•  ์ˆ˜ ์žˆ๋Š” ํ”Œ๋žซํผ์ž…๋‹ˆ๋‹ค. 2010๋…„์— Anthony Goldbloom๊ณผ Jeremy Howard์— ์˜ํ•ด ์‹œ์ž‘๋˜์–ด 2017๋…„์— Google ์— ์ธ์ˆ˜๋˜์—ˆ์Šต๋‹ˆ๋‹ค. Kaggle์„ ํ†ตํ•ด ์‚ฌ์šฉ์ž๋Š” ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์„ ์—ฐ๊ฒฐ, ๊ฒ€์ƒ‰ ๋ฐ ๊ณต์œ ํ•˜๊ณ , GPU-๊ธฐ๋ฐ˜ ๋…ธํŠธ๋ถ์„ ์‚ฌ์šฉํ•˜๋ฉฐ, ๋ฐ์ดํ„ฐ ๊ณผํ•™ ๊ฒฝ์—ฐ ๋Œ€ํšŒ์— ์ฐธ๊ฐ€ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ํ”Œ๋žซํผ์€ ๊ฐ•๋ ฅํ•œ ๋„๊ตฌ์™€ ๋ฆฌ์†Œ์Šค๋ฅผ ์ œ๊ณตํ•˜์—ฌ ๋…ธ๋ จํ•œ ์ „๋ฌธ๊ฐ€์™€ ์—ด์„ฑ์ ์ธ ํ•™์Šต์ž ๋ชจ๋‘๊ฐ€ ๋ชฉํ‘œ๋ฅผ ๋‹ฌ์„ฑํ•  ์ˆ˜ ์žˆ๋„๋ก ์„ค๊ณ„๋˜์—ˆ์Šต๋‹ˆ๋‹ค.

2022๋…„ ํ˜„์žฌ ์ฒœ๋งŒ ๋ช… ์ด์ƒ์˜ ์‚ฌ์šฉ์ž๋ฅผ ๋ณด์œ ํ•œ Kaggle์€ ๋จธ์‹  ๋Ÿฌ๋‹ ๋ชจ๋ธ์„ ๊ฐœ๋ฐœํ•˜๊ณ  ์‹คํ—˜ํ•  ์ˆ˜ ์žˆ๋Š” ํ’๋ถ€ํ•œ ํ™˜๊ฒฝ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค. ๋กœ์ปฌ ์ปดํ“จํ„ฐ์˜ ์‚ฌ์–‘์ด๋‚˜ ์„ค์ •์— ๋Œ€ํ•ด ๊ฑฑ์ •ํ•  ํ•„์š” ์—†์ด Kaggle ๊ณ„์ •๊ณผ ์›น ๋ธŒ๋ผ์šฐ์ €๋งŒ ์žˆ์œผ๋ฉด ๋ฐ”๋กœ ์‹œ์ž‘ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Training YOLO11 Using Kaggle

Training YOLO11 models on Kaggle is simple and efficient, thanks to the platform's access to powerful GPUs.

To get started, access the Kaggle YOLO11 Notebook. Kaggle's environment comes with pre-installed libraries like TensorFlow and PyTorch, making the setup process hassle-free.

What is the kaggle integration with respect to YOLO11?

Once you sign in to your Kaggle account, you can click on the option to copy and edit the code, select a GPU under the accelerator settings, and run the notebook's cells to begin training your model. For a detailed understanding of the model training process and best practices, refer to our YOLO11 Model Training guide.

๋จธ์‹  ๋Ÿฌ๋‹ ๋ชจ๋ธ ํ•™์Šต์„ ์œ„ํ•ด kaggle์„ ์‚ฌ์šฉํ•˜์—ฌ GPU

On the official YOLO11 Kaggle notebook page, if you click on the three dots in the upper right-hand corner, you'll notice more options will pop up.

Overview of Options From the Official YOLO11 Kaggle Notebook Page

์ด๋Ÿฌํ•œ ์˜ต์…˜์—๋Š” ๋‹ค์Œ์ด ํฌํ•จ๋ฉ๋‹ˆ๋‹ค:

  • ๋ฒ„์ „ ๋ณด๊ธฐ: ๋…ธํŠธ๋ถ์˜ ์—ฌ๋Ÿฌ ๋ฒ„์ „์„ ํƒ์ƒ‰ํ•ด ์‹œ๊ฐ„ ๊ฒฝ๊ณผ์— ๋”ฐ๋ฅธ ๋ณ€๊ฒฝ ์‚ฌํ•ญ์„ ํ™•์ธํ•˜๊ณ  ํ•„์š”ํ•œ ๊ฒฝ์šฐ ์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆด ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • API ๋ช…๋ น ๋ณต์‚ฌ: ์ž๋™ํ™” ๋ฐ ์›Œํฌํ”Œ๋กœ์šฐ ํ†ตํ•ฉ์— ์œ ์šฉํ•œ ๋…ธํŠธ๋ถ๊ณผ ํ”„๋กœ๊ทธ๋ž˜๋ฐ ๋ฐฉ์‹์œผ๋กœ ์ƒํ˜ธ ์ž‘์šฉํ•  ์ˆ˜ ์žˆ๋Š” API ๋ช…๋ น์„ ๊ฐ€์ ธ์˜ต๋‹ˆ๋‹ค.
  • Google ๋…ธํŠธ๋ถ์—์„œ ์—ฝ๋‹ˆ๋‹ค: Google ์˜ ํ˜ธ์ŠคํŒ…๋œ ๋…ธํŠธ๋ถ ํ™˜๊ฒฝ์—์„œ ๋…ธํŠธ๋ถ์„ ์—ฝ๋‹ˆ๋‹ค.
  • Colab์—์„œ ์—ฝ๋‹ˆ๋‹ค: Google Colab์—์„œ ๋…ธํŠธ๋ถ์„ ์‹คํ–‰ํ•˜์—ฌ ์ถ”๊ฐ€ ํŽธ์ง‘ ๋ฐ ์‹คํ–‰์„ ํ•˜์„ธ์š”.
  • ๋Œ“๊ธ€ ํŒ”๋กœ์šฐ: ๋Œ“๊ธ€ ์„น์…˜์„ ๊ตฌ๋…ํ•˜์—ฌ ์—…๋ฐ์ดํŠธ๋ฅผ ๋ฐ›๊ณ  ์ปค๋ฎค๋‹ˆํ‹ฐ์— ์ฐธ์—ฌํ•˜์„ธ์š”.
  • ์ฝ”๋“œ ๋‹ค์šด๋กœ๋“œ: ์˜คํ”„๋ผ์ธ์—์„œ ์‚ฌ์šฉํ•˜๊ฑฐ๋‚˜ ๋กœ์ปฌ ํ™˜๊ฒฝ์—์„œ ๋ฒ„์ „ ๊ด€๋ฆฌ๋ฅผ ์œ„ํ•ด ์ „์ฒด ๋…ธํŠธ๋ถ์„ Jupyter(.ipynb) ํŒŒ์ผ๋กœ ๋‹ค์šด๋กœ๋“œํ•˜์„ธ์š”.
  • ์ปฌ๋ ‰์…˜์— ์ถ”๊ฐ€: ์‰ฝ๊ฒŒ ์•ก์„ธ์Šคํ•˜๊ณ  ์ •๋ฆฌํ•  ์ˆ˜ ์žˆ๋„๋ก ๋…ธํŠธ๋ถ์„ Kaggle ๊ณ„์ • ๋‚ด ์ปฌ๋ ‰์…˜์— ์ €์žฅํ•˜์„ธ์š”.
  • ๋ถ๋งˆํฌ: ๋‚˜์ค‘์— ๋น ๋ฅด๊ฒŒ ์•ก์„ธ์Šคํ•  ์ˆ˜ ์žˆ๋„๋ก ๋…ธํŠธ๋ถ์„ ๋ถ๋งˆํฌ์— ์ถ”๊ฐ€ํ•˜์„ธ์š”.
  • ๋…ธํŠธ๋ถ ํผ๊ฐ€๊ธฐ: ๋ธ”๋กœ๊ทธ, ์›น์‚ฌ์ดํŠธ, ๋ฌธ์„œ์— ๋…ธํŠธ๋ถ์„ ํฌํ•จํ•  ์ˆ˜ ์žˆ๋Š” ์ž„๋ฒ ๋“œ ๋งํฌ๋ฅผ ๋ฐ›์œผ์„ธ์š”.

Kaggle๋กœ ์ž‘์—…ํ•  ๋•Œ ๋ฐœ์ƒํ•˜๋Š” ์ผ๋ฐ˜์ ์ธ ๋ฌธ์ œ

Kaggle๋กœ ์ž‘์—…ํ•  ๋•Œ ๋ช‡ ๊ฐ€์ง€ ์ผ๋ฐ˜์ ์ธ ๋ฌธ์ œ์— ์ง๋ฉดํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋‹ค์Œ์€ ํ”Œ๋žซํผ์„ ์›ํ™œํ•˜๊ฒŒ ํƒ์ƒ‰ํ•˜๋Š” ๋ฐ ๋„์›€์ด ๋˜๋Š” ๋ช‡ ๊ฐ€์ง€ ์‚ฌํ•ญ์ž…๋‹ˆ๋‹ค:

  • Access to GPUs: In your Kaggle notebooks, you can activate a GPU at any time, with usage allowed for up to 30 hours per week. Kaggle provides the NVIDIA Tesla P100 GPU with 16GB of memory and also offers the option of using a NVIDIA GPU T4 x2. Powerful hardware accelerates your machine-learning tasks, making model training and inference much faster.
  • Kaggle ์ปค๋„: Kaggle Kernels๋Š” GPU๋ฅผ ํ†ตํ•ฉํ•˜์—ฌ ํด๋ผ์šฐ๋“œ ์ปดํ“จํ„ฐ์—์„œ ๋จธ์‹  ๋Ÿฌ๋‹ ์ž‘์—…์„ ์ˆ˜ํ–‰ํ•  ์ˆ˜ ์žˆ๋Š” ๋ฌด๋ฃŒ Jupyter ๋…ธํŠธ๋ถ ์„œ๋ฒ„์ž…๋‹ˆ๋‹ค. ๋‚ด ์ปดํ“จํ„ฐ์˜ CPU ์— ์˜์กดํ•  ํ•„์š”๊ฐ€ ์—†์œผ๋ฏ€๋กœ ๊ณผ๋ถ€ํ•˜๋ฅผ ํ”ผํ•˜๊ณ  ๋กœ์ปฌ ๋ฆฌ์†Œ์Šค๋ฅผ ํ™•๋ณดํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • Kaggle ๋ฐ์ดํ„ฐ ์„ธํŠธ: Kaggle ๋ฐ์ดํ„ฐ ์„ธํŠธ๋Š” ๋ฌด๋ฃŒ๋กœ ๋‹ค์šด๋กœ๋“œํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ๊ฐ ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์˜ ๋ผ์ด์„ ์Šค๋ฅผ ํ™•์ธํ•˜์—ฌ ์‚ฌ์šฉ ์ œํ•œ ์‚ฌํ•ญ์„ ํŒŒ์•…ํ•˜๋Š” ๊ฒƒ์ด ์ค‘์š”ํ•ฉ๋‹ˆ๋‹ค. ์ผ๋ถ€ ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์—๋Š” ํ•™์ˆ  ์ถœํŒ ๋˜๋Š” ์ƒ์—…์  ์‚ฌ์šฉ์— ๋Œ€ํ•œ ์ œํ•œ์ด ์žˆ์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฐ์ดํ„ฐ ์„ธํŠธ๋Š” Kaggle ๋…ธํŠธ๋ถ์— ์ง์ ‘ ๋‹ค์šด๋กœ๋“œํ•˜๊ฑฐ๋‚˜ Kaggle API๋ฅผ ํ†ตํ•ด ๋‹ค๋ฅธ ๊ณณ์—์„œ ๋‹ค์šด๋กœ๋“œํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ๋…ธํŠธ๋ถ ์ €์žฅ ๋ฐ ์ปค๋ฐ‹ํ•˜๊ธฐ: Kaggle์—์„œ ๋…ธํŠธ๋ถ์„ ์ €์žฅํ•˜๊ณ  ์ปค๋ฐ‹ํ•˜๋ ค๋ฉด "๋ฒ„์ „ ์ €์žฅ"์„ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋ฉด ๋…ธํŠธ๋ถ์˜ ํ˜„์žฌ ์ƒํƒœ๊ฐ€ ์ €์žฅ๋ฉ๋‹ˆ๋‹ค. ๋ฐฑ๊ทธ๋ผ์šด๋“œ ์ปค๋„์ด ์ถœ๋ ฅ ํŒŒ์ผ ์ƒ์„ฑ์„ ์™„๋ฃŒํ•˜๋ฉด ๋ฉ”์ธ ๋…ธํŠธ๋ถ ํŽ˜์ด์ง€์˜ ์ถœ๋ ฅ ํƒญ์—์„œ ํ•ด๋‹น ํŒŒ์ผ์— ์•ก์„ธ์Šคํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ํ˜‘์—…: Kaggle์€ ๊ณต๋™ ์ž‘์—…์„ ์ง€์›ํ•˜์ง€๋งŒ ์—ฌ๋Ÿฌ ์‚ฌ์šฉ์ž๊ฐ€ ๋™์‹œ์— ๋…ธํŠธ๋ถ์„ ํŽธ์ง‘ํ•  ์ˆ˜๋Š” ์—†์Šต๋‹ˆ๋‹ค. Kaggle์˜ ํ˜‘์—…์€ ๋น„๋™๊ธฐ์‹์œผ๋กœ ์ด๋ฃจ์–ด์ง€๋ฏ€๋กœ ์‚ฌ์šฉ์ž๋“ค์€ ์„œ๋กœ ๋‹ค๋ฅธ ์‹œ๊ฐ„์— ๊ฐ™์€ ๋…ธํŠธ๋ถ์„ ๊ณต์œ ํ•˜๊ณ  ์ž‘์—…ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆฌ๊ธฐ: ๋…ธํŠธ๋ถ์˜ ์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆฌ๋ ค๋ฉด ๋…ธํŠธ๋ถ์„ ์—ด๊ณ  ์˜ค๋ฅธ์ชฝ ์ƒ๋‹จ ๋ชจ์„œ๋ฆฌ์— ์žˆ๋Š” ์„ธ๋กœ ์  3๊ฐœ๋ฅผ ํด๋ฆญํ•ด "๋ฒ„์ „ ๋ณด๊ธฐ"๋ฅผ ์„ ํƒํ•ฉ๋‹ˆ๋‹ค. ๋˜๋Œ๋ฆฌ๊ณ  ์‹ถ์€ ๋ฒ„์ „์„ ์ฐพ์•„ ์˜†์— ์žˆ๋Š” "..." ๋ฉ”๋‰ด๋ฅผ ํด๋ฆญํ•˜๊ณ  "๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆฌ๊ธฐ"๋ฅผ ์„ ํƒํ•ฉ๋‹ˆ๋‹ค. ๋…ธํŠธ๋ถ์ด ๋˜๋Œ๋ฆฌ๋ฉด "๋ฒ„์ „ ์ €์žฅ"์„ ํด๋ฆญํ•ด ๋ณ€๊ฒฝ ์‚ฌํ•ญ์„ ์ปค๋ฐ‹ํ•ฉ๋‹ˆ๋‹ค.

Kaggle์˜ ์ฃผ์š” ๊ธฐ๋Šฅ

๋‹ค์Œ์œผ๋กœ, ๋ฐ์ดํ„ฐ ๊ณผํ•™ ๋ฐ ๋จธ์‹  ๋Ÿฌ๋‹ ์• ํ˜ธ๊ฐ€๋“ค์„ ์œ„ํ•œ ํ›Œ๋ฅญํ•œ ํ”Œ๋žซํผ์ด ๋  ์ˆ˜ ์žˆ๋Š” Kaggle์˜ ๊ธฐ๋Šฅ์— ๋Œ€ํ•ด ์•Œ์•„๋ด…์‹œ๋‹ค. ๋‹ค์Œ์€ ๋ช‡ ๊ฐ€์ง€ ์ฃผ์š” ํŠน์ง•์ž…๋‹ˆ๋‹ค:

  • Datasets: Kaggle hosts a massive collection of datasets on various topics. You can easily search and use these datasets in your projects, which is particularly handy for training and testing your YOLO11 models.
  • ๋Œ€ํšŒ: ํฅ๋ฏธ์ง„์ง„ํ•œ ๋Œ€ํšŒ๋กœ ์ž˜ ์•Œ๋ ค์ง„ Kaggle์€ ๋ฐ์ดํ„ฐ ๊ณผํ•™์ž์™€ ๋จธ์‹  ๋Ÿฌ๋‹ ์• ํ˜ธ๊ฐ€๋“ค์ด ์‹ค์ œ ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•  ์ˆ˜ ์žˆ๊ฒŒ ํ•ด์ค๋‹ˆ๋‹ค. ๊ฒฝ์Ÿ์„ ํ†ตํ•ด ๊ธฐ์ˆ ์„ ํ–ฅ์ƒํ•˜๊ณ , ์ƒˆ๋กœ์šด ๊ธฐ์ˆ ์„ ๋ฐฐ์šฐ๊ณ , ์ปค๋ฎค๋‹ˆํ‹ฐ์—์„œ ์ธ์ •์„ ๋ฐ›์„ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • Free Access to TPUs: Kaggle provides free access to powerful TPUs, which are essential for training complex machine learning models. This means you can speed up processing and boost the performance of your YOLO11 projects without incurring extra costs.
  • Github๊ณผ ํ†ตํ•ฉ: Kaggle์„ ์‚ฌ์šฉํ•˜๋ฉด GitHub ์ €์žฅ์†Œ๋ฅผ ๊ฐ„ํŽธํ•˜๊ฒŒ ์—ฐ๊ฒฐํ•ด ๋…ธํŠธ๋ถ์„ ์—…๋กœ๋“œํ•˜๊ณ  ์ž‘์—…์„ ์ €์žฅํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ด ํ†ตํ•ฉ ๊ธฐ๋Šฅ์œผ๋กœ ํŒŒ์ผ์„ ํŽธ๋ฆฌํ•˜๊ฒŒ ๊ด€๋ฆฌํ•˜๊ณ  ์•ก์„ธ์Šคํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ์ปค๋ฎค๋‹ˆํ‹ฐ ๋ฐ ํ† ๋ก : Kaggle์€ ๋ฐ์ดํ„ฐ ๊ณผํ•™์ž ๋ฐ ๋จธ์‹  ๋Ÿฌ๋‹ ์‹ค๋ฌด์ž๋กœ ๊ตฌ์„ฑ๋œ ๊ฐ•๋ ฅํ•œ ์ปค๋ฎค๋‹ˆํ‹ฐ๋ฅผ ์ž๋ž‘ํ•ฉ๋‹ˆ๋‹ค. ํ† ๋ก  ํฌ๋Ÿผ๊ณผ ๊ณต์œ  ๋…ธํŠธ๋ถ์€ ํ•™์Šต๊ณผ ๋ฌธ์ œ ํ•ด๊ฒฐ์„ ์œ„ํ•œ ํ›Œ๋ฅญํ•œ ๋ฆฌ์†Œ์Šค์ž…๋‹ˆ๋‹ค. ์‰ฝ๊ฒŒ ๋„์›€์„ ์ฐพ๊ณ , ์ง€์‹์„ ๊ณต์œ ํ•˜๊ณ , ๋‹ค๋ฅธ ์‚ฌ๋žŒ๋“ค๊ณผ ํ˜‘์—…ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Why Should You Use Kaggle for Your YOLO11 Projects?

๋จธ์‹  ๋Ÿฌ๋‹ ๋ชจ๋ธ์„ ํ›ˆ๋ จํ•˜๊ณ  ํ‰๊ฐ€ํ•˜๋Š” ๋ฐ๋Š” ์—ฌ๋Ÿฌ ํ”Œ๋žซํผ์ด ์žˆ๋Š”๋ฐ, Kaggle์ด ๋‹๋ณด์ด๋Š” ์ด์œ ๋Š” ๋ฌด์—‡์ผ๊นŒ์š”? ๋จธ์‹  ๋Ÿฌ๋‹ ํ”„๋กœ์ ํŠธ์— Kaggle์„ ์‚ฌ์šฉํ•˜๋ฉด ์–ป์„ ์ˆ˜ ์žˆ๋Š” ์ด์ ์— ๋Œ€ํ•ด ์ž์„ธํžˆ ์•Œ์•„๋ณด์„ธ์š”:

  • Public Notebooks: You can make your Kaggle notebooks public, allowing other users to view, vote, fork, and discuss your work. Kaggle promotes collaboration, feedback, and the sharing of ideas, helping you improve your YOLO11 models.
  • ๋…ธํŠธ๋ถ ์ปค๋ฐ‹์˜ ์ข…ํ•ฉ์ ์ธ ํžˆ์Šคํ† ๋ฆฌ: Kaggle์€ ๋…ธํŠธ๋ถ ์ปค๋ฐ‹์— ๋Œ€ํ•œ ์ž์„ธํ•œ ํžˆ์Šคํ† ๋ฆฌ๋ฅผ ์ƒ์„ฑํ•ฉ๋‹ˆ๋‹ค. ์ด๋ฅผ ํ†ตํ•ด ์‹œ๊ฐ„ ๊ฒฝ๊ณผ์— ๋”ฐ๋ฅธ ๋ณ€๊ฒฝ ์‚ฌํ•ญ์„ ๊ฒ€ํ† ํ•˜๊ณ  ์ถ”์ ํ•  ์ˆ˜ ์žˆ์–ด ํ”„๋กœ์ ํŠธ์˜ ์ง„ํ™”๋ฅผ ๋” ์‰ฝ๊ฒŒ ์ดํ•ดํ•˜๊ณ  ํ•„์š”ํ•œ ๊ฒฝ์šฐ ์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆด ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ์ฝ˜์†” ์•ก์„ธ์Šค: Kaggle์€ ์ฝ˜์†”์„ ์ œ๊ณตํ•˜์—ฌ ํ™˜๊ฒฝ์„ ๋” ์ž˜ ์ œ์–ดํ•  ์ˆ˜ ์žˆ๋„๋ก ํ•ด์ค๋‹ˆ๋‹ค. ์ด ๊ธฐ๋Šฅ์„ ์‚ฌ์šฉํ•˜๋ฉด ๋ช…๋ น์ค„์—์„œ ์ง์ ‘ ๋‹ค์–‘ํ•œ ์ž‘์—…์„ ์ˆ˜ํ–‰ํ•˜์—ฌ ์›Œํฌํ”Œ๋กœ์™€ ์ƒ์‚ฐ์„ฑ์„ ํ–ฅ์ƒ์‹œํ‚ฌ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ๋ฆฌ์†Œ์Šค ๊ฐ€์šฉ์„ฑ: Kaggle์˜ ๊ฐ ๋…ธํŠธ๋ถ ํŽธ์ง‘ ์„ธ์…˜์—๋Š” ์ƒ๋‹นํ•œ ๋ฆฌ์†Œ์Šค๊ฐ€ ์ œ๊ณต๋ฉ๋‹ˆ๋‹ค: CPU ๋ฐ GPU ์„ธ์…˜์˜ ์‹คํ–‰ ์‹œ๊ฐ„์€ 12์‹œ๊ฐ„, TPU ์„ธ์…˜์˜ ์‹คํ–‰ ์‹œ๊ฐ„์€ 9์‹œ๊ฐ„, ์ž๋™ ์ €์žฅ๋œ ๋””์Šคํฌ ๊ณต๊ฐ„์€ 20๊ธฐ๊ฐ€๋ฐ”์ดํŠธ์ž…๋‹ˆ๋‹ค.
  • ๋…ธํŠธ๋ถ ์˜ˆ์•ฝ: Kaggle์„ ์‚ฌ์šฉํ•˜๋ฉด ๋…ธํŠธ๋ถ์„ ํŠน์ • ์‹œ๊ฐ„์— ์‹คํ–‰ํ•˜๋„๋ก ์˜ˆ์•ฝํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์ผ์ •ํ•œ ๊ฐ„๊ฒฉ์œผ๋กœ ๋ชจ๋ธ์„ ํ›ˆ๋ จํ•˜๋Š” ๊ฒƒ๊ณผ ๊ฐ™์ด ์ˆ˜๋™ ๊ฐœ์ž… ์—†์ด ๋ฐ˜๋ณต์ ์ธ ์ž‘์—…์„ ์ž๋™ํ™”ํ•  ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

Kaggle์— ๋Œ€ํ•ด ๊ณ„์† ์•Œ์•„๋ณด๊ธฐ

Kaggle์— ๋Œ€ํ•ด ์ž์„ธํžˆ ์•Œ์•„๋ณด๋ ค๋ฉด ๋‹ค์Œ๊ณผ ๊ฐ™์€ ์œ ์šฉํ•œ ๋ฆฌ์†Œ์Šค๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”:

  • ์บ๊ธ€ ํ•™์Šต: Kaggle Learn์—์„œ ๋‹ค์–‘ํ•œ ๋ฌด๋ฃŒ ๋Œ€ํ™”ํ˜• ์ž์Šต์„œ๋ฅผ ์ฐพ์•„๋ณด์„ธ์š”. ์ด๋Ÿฌํ•œ ๊ต์œก ๊ณผ์ •์—์„œ๋Š” ํ•„์ˆ˜ ๋ฐ์ดํ„ฐ ๊ณผํ•™ ์ฃผ์ œ๋ฅผ ๋‹ค๋ฃจ๊ณ  ์ƒˆ๋กœ์šด ๊ธฐ์ˆ ์„ ์Šต๋“ํ•˜๋Š” ๋ฐ ๋„์›€์ด ๋˜๋Š” ์‹ค์Šต ๊ฒฝํ—˜์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค.
  • Kaggle ์‹œ์ž‘ํ•˜๊ธฐ: ์ด ์ข…ํ•ฉ ๊ฐ€์ด๋“œ๋Š” ๋Œ€ํšŒ ์ฐธ๊ฐ€๋ถ€ํ„ฐ ์ฒซ ๋…ธํŠธ๋ถ ๋งŒ๋“ค๊ธฐ๊นŒ์ง€ Kaggle ์‚ฌ์šฉ์˜ ๊ธฐ๋ณธ ์‚ฌํ•ญ์„ ์•ˆ๋‚ดํ•ฉ๋‹ˆ๋‹ค. ์ฒ˜์Œ ์‹œ์ž‘ํ•˜๋Š” ๋ถ„๋“ค์—๊ฒŒ ์ข‹์€ ์ถœ๋ฐœ์ ์ด ๋  ๊ฒƒ์ž…๋‹ˆ๋‹ค.
  • ์บ๊ธ€ ๋ฏธ๋””์—„ ํŽ˜์ด์ง€: Kaggle์˜ Medium ํŽ˜์ด์ง€์—์„œ ํŠœํ† ๋ฆฌ์–ผ, ์—…๋ฐ์ดํŠธ ๋ฐ ์ปค๋ฎค๋‹ˆํ‹ฐ ๊ธฐ์—ฌ๋ฅผ ์‚ดํŽด๋ณด์„ธ์š”. ์ตœ์‹  ๋™ํ–ฅ์„ ํŒŒ์•…ํ•˜๊ณ  ๋ฐ์ดํ„ฐ ๊ณผํ•™์— ๋Œ€ํ•œ ๋” ๊นŠ์€ ์ธ์‚ฌ์ดํŠธ๋ฅผ ์–ป์„ ์ˆ˜ ์žˆ๋Š” ํ›Œ๋ฅญํ•œ ์ž๋ฃŒ์ž…๋‹ˆ๋‹ค.

์š”์•ฝ

We've seen how Kaggle can boost your YOLO11 projects by providing free access to powerful GPUs, making model training and evaluation efficient. Kaggle's platform is user-friendly, with pre-installed libraries for quick setup.

์ž์„ธํ•œ ๋‚ด์šฉ์€ Kaggle์˜ ์„ค๋ช…์„œ๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”.

Interested in more YOLO11 integrations? Check out the Ultralytics integration guide to explore additional tools and capabilities for your machine learning projects.

์ž์ฃผ ๋ฌป๋Š” ์งˆ๋ฌธ

How do I train a YOLO11 model on Kaggle?

Training a YOLO11 model on Kaggle is straightforward. First, access the Kaggle YOLO11 Notebook. Sign in to your Kaggle account, copy and edit the notebook, and select a GPU under the accelerator settings. Run the notebook cells to start training. For more detailed steps, refer to our YOLO11 Model Training guide.

What are the benefits of using Kaggle for YOLO11 model training?

Kaggle offers several advantages for training YOLO11 models:

  • Free GPU Access: Utilize powerful GPUs like NVIDIA Tesla P100 or T4 x2 for up to 30 hours per week.
  • ์‚ฌ์ „ ์„ค์น˜๋œ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ: TensorFlow ๋ฐ PyTorch ๊ฐ™์€ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๊ฐ€ ์‚ฌ์ „ ์„ค์น˜๋˜์–ด ์žˆ์–ด ์„ค์ •์ด ๊ฐ„ํŽธํ•ฉ๋‹ˆ๋‹ค.
  • ์ปค๋ฎค๋‹ˆํ‹ฐ ํ˜‘์—…: ๋ฐฉ๋Œ€ํ•œ ๋ฐ์ดํ„ฐ ๊ณผํ•™์ž ๋ฐ ๋จธ์‹  ๋Ÿฌ๋‹ ์• ํ˜ธ๊ฐ€๋กœ ๊ตฌ์„ฑ๋œ ์ปค๋ฎค๋‹ˆํ‹ฐ์— ์ฐธ์—ฌํ•˜์„ธ์š”.
  • ๋ฒ„์ „ ๊ด€๋ฆฌ: ๋…ธํŠธ๋ถ์˜ ์—ฌ๋Ÿฌ ๋ฒ„์ „์„ ๊ฐ„ํŽธํ•˜๊ฒŒ ๊ด€๋ฆฌํ•˜๊ณ  ํ•„์š”ํ•œ ๊ฒฝ์šฐ ์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆด ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.

์ž์„ธํ•œ ๋‚ด์šฉ์€ Ultralytics ํ†ตํ•ฉ ๊ฐ€์ด๋“œ๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”.

What common issues might I encounter when using Kaggle for YOLO11, and how can I resolve them?

์ผ๋ฐ˜์ ์ธ ๋ฌธ์ œ๋Š” ๋‹ค์Œ๊ณผ ๊ฐ™์Šต๋‹ˆ๋‹ค:

  • GPU์— ์•ก์„ธ์Šคํ•ฉ๋‹ˆ๋‹ค: ๋…ธํŠธ๋ถ ์„ค์ •์—์„œ GPU ์„ ํ™œ์„ฑํ™”ํ–ˆ๋Š”์ง€ ํ™•์ธํ•˜์„ธ์š”. Kaggle์€ ์ฃผ๋‹น ์ตœ๋Œ€ 30์‹œ๊ฐ„์˜ GPU ์‚ฌ์šฉ์„ ํ—ˆ์šฉํ•ฉ๋‹ˆ๋‹ค.
  • ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ ๋ผ์ด์„ ์Šค: ๊ฐ ๋ฐ์ดํ„ฐ ์ง‘ํ•ฉ์˜ ๋ผ์ด์„ ์Šค๋ฅผ ํ™•์ธํ•˜์—ฌ ์‚ฌ์šฉ ์ œํ•œ ์‚ฌํ•ญ์„ ํŒŒ์•…ํ•˜์„ธ์š”.
  • ๋…ธํŠธ๋ถ ์ €์žฅ ๋ฐ ์ปค๋ฐ‹: "๋ฒ„์ „ ์ €์žฅ"์„ ํด๋ฆญํ•ด ๋…ธํŠธ๋ถ์˜ ์ƒํƒœ๋ฅผ ์ €์žฅํ•˜๊ณ  ์ถœ๋ ฅ ํƒญ์—์„œ ์ถœ๋ ฅ ํŒŒ์ผ์— ์•ก์„ธ์Šคํ•ฉ๋‹ˆ๋‹ค.
  • ํ˜‘์—…: ์—ฌ๋Ÿฌ ์‚ฌ์šฉ์ž๊ฐ€ ๋™์‹œ์— ๋…ธํŠธ๋ถ์„ ํŽธ์ง‘ํ•  ์ˆ˜ ์—†๋Š” ๋น„๋™๊ธฐ์‹ ๊ณต๋™ ์ž‘์—…์„ ์ง€์›ํ•ฉ๋‹ˆ๋‹ค.

๋” ๋งŽ์€ ๋ฌธ์ œ ํ•ด๊ฒฐ ํŒ์€ ์ผ๋ฐ˜์ ์ธ ๋ฌธ์ œ ๊ฐ€์ด๋“œ๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”.

Why should I choose Kaggle over other platforms like Google Colab for training YOLO11 models?

Kaggle์€ ํƒ์›”ํ•œ ์„ ํƒ์ด ๋  ์ˆ˜ ์žˆ๋Š” ๊ณ ์œ ํ•œ ๊ธฐ๋Šฅ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค:

  • ๊ณต๊ฐœ ๋…ธํŠธ๋ถ: ์ปค๋ฎค๋‹ˆํ‹ฐ์™€ ์ž‘์—…์„ ๊ณต์œ ํ•˜์—ฌ ํ”ผ๋“œ๋ฐฑ์„ ๋ฐ›๊ณ  ํ˜‘์—…ํ•˜์„ธ์š”.
  • TPU ๋ฌด๋ฃŒ ์ด์šฉ: ์ถ”๊ฐ€ ๋น„์šฉ ์—†์ด ๊ฐ•๋ ฅํ•œ TPU๋กœ ํ›ˆ๋ จ ์†๋„๋ฅผ ๋†’์ผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค.
  • ์ข…ํ•ฉ์ ์ธ ํžˆ์Šคํ† ๋ฆฌ: ๋…ธํŠธ๋ถ ์ปค๋ฐ‹์˜ ์ž์„ธํ•œ ํžˆ์Šคํ† ๋ฆฌ๋ฅผ ํ†ตํ•ด ์‹œ๊ฐ„ ๊ฒฝ๊ณผ์— ๋”ฐ๋ฅธ ๋ณ€๊ฒฝ ์‚ฌํ•ญ์„ ์ถ”์ ํ•˜์„ธ์š”.
  • ๋ฆฌ์†Œ์Šค ๊ฐ€์šฉ์„ฑ: CPU ๋ฐ GPU ์„ธ์…˜์˜ ๊ฒฝ์šฐ 12์‹œ๊ฐ„์˜ ์‹คํ–‰ ์‹œ๊ฐ„์„ ํฌํ•จํ•ด ๊ฐ ๋…ธํŠธ๋ถ ์„ธ์…˜์— ์ƒ๋‹นํ•œ ๋ฆฌ์†Œ์Šค๊ฐ€ ์ œ๊ณต๋ฉ๋‹ˆ๋‹ค. Google Colab๊ณผ ๋น„๊ตํ•˜๋ ค๋ฉด Google Colab ๊ฐ€์ด๋“œ๋ฅผ ์ฐธ์กฐํ•˜์„ธ์š”.

Kaggle ๋…ธํŠธ๋ถ์˜ ์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆฌ๋ ค๋ฉด ์–ด๋–ป๊ฒŒ ํ•ด์•ผ ํ•˜๋‚˜์š”?

์ด์ „ ๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆฌ๋ ค๋ฉด

  1. ๋…ธํŠธ๋ถ์„ ์—ด๊ณ  ์˜ค๋ฅธ์ชฝ ์ƒ๋‹จ ๋ชจ์„œ๋ฆฌ์— ์žˆ๋Š” ์„ธ๋กœ ์  3๊ฐœ๋ฅผ ํด๋ฆญํ•ฉ๋‹ˆ๋‹ค.
  2. "๋ฒ„์ „ ๋ณด๊ธฐ"๋ฅผ ์„ ํƒํ•ฉ๋‹ˆ๋‹ค.
  3. ๋˜๋Œ๋ฆฌ๋ ค๋Š” ๋ฒ„์ „์„ ์ฐพ์•„ ์˜†์— ์žˆ๋Š” '...' ๋ฉ”๋‰ด๋ฅผ ํด๋ฆญํ•œ ๋‹ค์Œ "๋ฒ„์ „์œผ๋กœ ๋˜๋Œ๋ฆฌ๊ธฐ"๋ฅผ ์„ ํƒํ•ฉ๋‹ˆ๋‹ค.
  4. '๋ฒ„์ „ ์ €์žฅ'์„ ํด๋ฆญํ•˜์—ฌ ๋ณ€๊ฒฝ ์‚ฌํ•ญ์„ ์ปค๋ฐ‹ํ•ฉ๋‹ˆ๋‹ค.

๐Ÿ“… Created 2 months ago โœ๏ธ Updated 2 days ago

๋Œ“๊ธ€