YOLOv10 Custom Dataset Training Object Detection Brain Tumors
HTML-код
- Опубликовано: 22 июл 2024
- I will show you how to train a custom model using YOLOv10 on a custom dataset and run inference. We will be using the custom dataset containing brain MRI scans from kaggle to predict positive and negative tumor locations.
0:00 Introduction
0:47 YOLOv10 Overview
2:37 Setup and Verify GPU
2:56 Install YOLOv10
4:02 Download YOLOv10 Weights
4:37 Prepare Dataset
6:17 Train YOLOv10
7:12 Review YOLOv10 Training Results
9:34 Detect Object using YOLOv10
Thanks for watching! If you found this video helpful, please like, subscribe and share:
/ @kevinwoodrobotics
Social:
LinkedIn: / kevinwoodrobotics
Github: github.com/kevinwoodrobotics
Instagram: / kevinwoodrobotics
Twitter: / kevinwoodrobotics
Notebook: kevinwoodrobotics.com/product/yolov10-custom-dataset-training-object-detection-brain-tumors/
Full Playlist Code and Doc: kevinwoodrobotics.com/product/ai-and-machine-learning-full-playlist/
Again nice one
Hello Sir, could you please explain what 'train/box_om', 'train/cls_om', 'train/dfl_om', 'train/box_oo', 'train/cls_oo', and 'train/dfl_oo' mean in the plot after the YOLOv10 training is completed?
train/box_om:
box_om stands for "box objectness metric."
It represents the objectness score for bounding boxes. This score indicates how confident the model is that a bounding box contains an object.
train/cls_om:
cls_om stands for "class objectness metric."
It represents the objectness score for class predictions. This score indicates how confident the model is about the class prediction for a detected object.
train/dfl_om:
dfl_om stands for "distribution focal loss objectness metric."
It represents the objectness score related to the distribution focal loss, which is a loss function used to refine the localization of bounding boxes.
train/box_oo:
box_oo stands for "box objectness object."
It represents the actual objectness score for bounding boxes, indicating the model's confidence in the presence of an object within the predicted boxes during training.
train/cls_oo:
cls_oo stands for "class objectness object."
It represents the actual objectness score for class predictions, indicating the model's confidence in the class of the detected object during training.
train/dfl_oo:
dfl_oo stands for "distribution focal loss objectness object."
It represents the actual objectness score related to the distribution focal loss during training, indicating how well the model is refining the bounding box localization.
When I use my custom dataset, I'm getting a error, could you help me?
What error?
When trying to validate, I am getting an error. Are you able to validate the results on a test set?
You can adjust the path to a test image using the predict command
where u get that notebook
kevinwoodrobotics.com/product/yolov10-custom-dataset-training-object-detection-brain-tumors/