Can i train YOLOV5 in video datset?

vijay · March 31, 2022, 9:53am

Hey,
Quick question!!. I have VisDrone Video dataset for object detection. Anyway i’ll use the model for video detection only.

Can i provide video dataset as input for training yolov5. If yes, how to do that? can someone provide reference for that.

Glenn · March 31, 2022, 11:12am

@vijay VisDone training is super simple. Dataset downloads automatically on first use:

python train.py --data VisDrone.yaml

github.com

ultralytics/yolov5/blob/c94736acece384ed2d5a7299ee82af2969abb48b/data/VisDrone.yaml#L1-L19

      
        
            # YOLOv5 🚀 by Ultralytics, GPL-3.0 license
            # VisDrone2019-DET dataset https://github.com/VisDrone/VisDrone-Dataset by Tianjin University
            # Example usage: python train.py --data VisDrone.yaml
            # parent
            # ├── yolov5
            # └── datasets
            #     └── VisDrone  ← downloads here
            
            

            
# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
            path: ../datasets/VisDrone  # dataset root dir
            train: VisDrone2019-DET-train/images  # train images (relative to 'path')  6471 images
            val: VisDrone2019-DET-val/images  # val images (relative to 'path')  548 images
            test: VisDrone2019-DET-test-dev/images  # test images (optional)  1610 images
            
            
# Classes
            nc: 10  # number of classes
            names: ['pedestrian', 'people', 'bicycle', 'car', 'van', 'truck', 'tricycle', 'awning-tricycle', 'bus', 'motor']

vijay · April 1, 2022, 9:09am

Thats awesome glenn.
Can i use this VisDrone-VID Dataset for training? It has video sequences not images. Even the labels files are in different files

Glenn · April 4, 2022, 11:17pm

@vijay oh! Well, you could parse the videos into images and train that way, but no, there’s no direct training on videos currently with YOLOv5.

Idea_Squad_Drone · October 16, 2023, 8:12pm

@Glenn I have a VisDrone-MOT dataset which consists of videos and its corresponding labels.
Should I now convert the videos into frames for creating the dataset(create yaml files too) or is there any way Yolov8 can be trained on Video dataset?