CS2YOLOv Project

This project is based on Ape-xCV/Apex-CV-YOLO-v8-Aim-Assist-Bot?tab=readme-ov-file

Project Description

The CS2YOLOv project focuses on developing and integrating various tools for working with the YOLO model specifically tailored for Counter-Strike 2 (CS2). This includes scripts for training the YOLO model, data augmentation, and automating data collection for training purposes. Each script serves a specific task, whether it's preparing datasets, augmenting images, or training the model.

Currently, the project is optimized for 4k resolution and a scaling factor of 200.

Main Project Scripts

splitDatasetFiles.py

A script for evenly distributing images and labels for training. Prepares training, validation, and test datasets by splitting images and corresponding labels into parts.

Usage:

python splitDatasetFiles.py

Albumentations.py:

This script performs the task of augmenting images and corresponding object labels (in YOLO format). It also removes images and labels with incorrect classes, improving data quality.

Main Functions:

• Reading images and labels.
• Applying augmentations such as random rotations, mirror flips, and shifts.
• Removing files with incorrect labels.

Usage:

python Albumentations.py

main.py:

This script implements an auto-targeting system based on the YOLO model, capturing frames from the screen and using a random cursor movement technique to target objects.

Main Functions:

• Screen frame capturing.
• Object detection using YOLO.
• Automatic targeting.

Usage:

python ApexBot.py

labelConfig.py:

A script for interactive editing of labels in YOLO format. Allows viewing images and corresponding labels, adding new labels, or removing existing ones.

Main Functions:

• Editing object labels on images.
• Adding and deleting labels.
• Removing images and labels with hotkeys.

Usage:

python labelConfig.py

semiauto_dataset_collector.py:

A system for automatically capturing and saving images using the YOLO model, which also displays detection results in real-time.

Main Functions:

• Screen video capture.
• YOLO processing for object detection.
• Saving images and annotations in YOLO format.

Usage:

python semiauto_dataset_collector.py

TrainYolov10.py:

A script for training the YOLO model using specified parameters. Logs results to TensorBoard and saves training graphs.

Main Functions:

• Training the YOLO model on a custom dataset.
• Logging metrics and training graphs to TensorBoard.
• Plotting graphs for analysis.

Usage:

python TrainYolov10.py

Installation

1. Clone the project repository or download the archive:

git clone https://github.com/your-repository/ApexYOLO.git

2. Install the required dependencies, including PyTorch and YOLO:

pip install -r requirements.txt
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

3. Additionally, install Albumentations for data augmentation:

pip install albumentations

Usage

To execute each script, follow the description above. For example, to train the YOLO model, run:

python TrainYolov10.py

To start the auto-targeting system:

python ApexBot.py

Logs and TensorBoard

To monitor the training process, use TensorBoard. To start TensorBoard and view the training metrics, execute:

tensorboard --logdir=runs/yolo_training10/ApexEsp80

Project Structure

• CS2YOLO/ — Root project folder.
• dataSet/ — Dataset used for training and testing.
• runs/ — Folder for saving logs, metrics, and training results.
• scripts/ — Folder containing the main project scripts.