WesleyH101
Voice Search Optimizer
Divine
2
MONTHS
2 2 MONTHS OF SERVICE
LEVEL 1
300 XP
Today I'm going to teach you how to make AI aimbot for almost any game, with a simple python source and Yolo v5.
STEP 1:
First you can start by getting the script from the link below:
After we have downloaded the source, we need to download all dependencies to run the script. You can find them in the requirments.txt file included in the source.
STEP 2:
Now that you have the dependencies installed to run the script, you will need to annotate images for the game manually, or you can save time by searching the game in the link below to find images already annotated.
Once you have found a Dataset that you feel confident in, download it. ( Annotations with body and head will make aimbot better)
STEP 3:
Now that we have our Dataset we will need to "train" our CPU to be able to detect enemies in game real time.
to do that we will be using yolov5. (this is the yolo version that works with script currently, yolov8 does not work with this script)
To train model via Yolov5, Open up Google Collab.
https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb#scrollTo=ZY2VXXXu74w5
Click the folder on the left side, and drag your dataset to folder.
[img]
After that's done, connect a google drive to save trained model , and install YoloV5 dependencies on the collab site under "setup".
IF YOU MANUALLY ANNOTED DATASET:
You will need to modify the "coco128" file in the Yolov5 folder. In most cases you only annotate head and body for aimbot so you need to make sure to label those. Edit the coco128 file like so:
[img]
IF YOU DOWNLOADED PREMADE DATASET:
Some datasets include a Yaml file , if that's the case, delete the coco128.YAML file and replace with YAML from dataset.
To unzip dataset folder once in the directory, you can create or add code and use:
!unzip -q /content/dataset.zip -d../ and run code.
You can create an "output" folder in your google drive to store trained models.
Now its time to train. [img]
you can skip step 1 and 2 in the collab tutorial because we already detected figures.
Go to step 3.
Train at image size 640, batch 32, and 300 epochs. It should look like this:
Code:
[code]
!python train.py --img 640 --batch 32 --epochs 300 --data coco128.yaml --weights yolov5s.pt --cache --project
[/code]
If you are using your downloaded dataset simply replace "coco128.yaml" with downloaded dataset Yaml file. or replace coco128 contents with contents of YAML file from downloaded dataset.
Press Train. This will take some time.
Congratulations we are through the "hard" part.
STEP 4
Now that we have trained our model you will need to go to your google drive and find a file called "best" , this will be the best trained model and what you will need to implement into the source to make sure it works with the game you trained and are trying to make aimbot for.
In the "Lib" folder of the source , you will find two files. Aimbot and Best. Replace the best with the new best from your google drive.
There we have it .
CONCLUSION:
This source is undetected by most major Anti cheats, and works well. If you are experienced with Python you can tweak the source to your liking.
Whenever you want to change games just re-train model for that game and replace the "best" file.
STEP 1:
First you can start by getting the script from the link below:
You must upgrade your account or reply in the thread to view hidden text.
After we have downloaded the source, we need to download all dependencies to run the script. You can find them in the requirments.txt file included in the source.
STEP 2:
Now that you have the dependencies installed to run the script, you will need to annotate images for the game manually, or you can save time by searching the game in the link below to find images already annotated.
Once you have found a Dataset that you feel confident in, download it. ( Annotations with body and head will make aimbot better)
STEP 3:
Now that we have our Dataset we will need to "train" our CPU to be able to detect enemies in game real time.
to do that we will be using yolov5. (this is the yolo version that works with script currently, yolov8 does not work with this script)
To train model via Yolov5, Open up Google Collab.
https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb#scrollTo=ZY2VXXXu74w5
Click the folder on the left side, and drag your dataset to folder.
[img]
After that's done, connect a google drive to save trained model , and install YoloV5 dependencies on the collab site under "setup".
IF YOU MANUALLY ANNOTED DATASET:
You will need to modify the "coco128" file in the Yolov5 folder. In most cases you only annotate head and body for aimbot so you need to make sure to label those. Edit the coco128 file like so:
[img]
IF YOU DOWNLOADED PREMADE DATASET:
Some datasets include a Yaml file , if that's the case, delete the coco128.YAML file and replace with YAML from dataset.
To unzip dataset folder once in the directory, you can create or add code and use:
!unzip -q /content/dataset.zip -d../ and run code.
You can create an "output" folder in your google drive to store trained models.
Now its time to train. [img]
you can skip step 1 and 2 in the collab tutorial because we already detected figures.
Go to step 3.
Train at image size 640, batch 32, and 300 epochs. It should look like this:
Code:
[code]
!python train.py --img 640 --batch 32 --epochs 300 --data coco128.yaml --weights yolov5s.pt --cache --project
[/code]
If you are using your downloaded dataset simply replace "coco128.yaml" with downloaded dataset Yaml file. or replace coco128 contents with contents of YAML file from downloaded dataset.
Press Train. This will take some time.
Congratulations we are through the "hard" part.
STEP 4
Now that we have trained our model you will need to go to your google drive and find a file called "best" , this will be the best trained model and what you will need to implement into the source to make sure it works with the game you trained and are trying to make aimbot for.
In the "Lib" folder of the source , you will find two files. Aimbot and Best. Replace the best with the new best from your google drive.
There we have it .
CONCLUSION:
This source is undetected by most major Anti cheats, and works well. If you are experienced with Python you can tweak the source to your liking.
Whenever you want to change games just re-train model for that game and replace the "best" file.