YOLOv11n-Training-Automation

This is meant for Windows only and not tested for Mac or Linux

Required: Python 3.12 Required for GPU Training

Python 3.13+ does NOT support CUDA PyTorch yet Check your Python Version

python --version

Steps for running

• Install requirements

pip install -r requirements.txt

• Add your Google API Key to .env file
• Verify prompts.csv file to ensure correct prompts
• Start by running prompt_automation.py
• After prompts are generated run visualize.py to check labels for each image. Any mistakes may need to be annotated manually
• Run organize_dataset.py to split annotated frames up into 70%, 20%, 10% for training, validations, and testing
• After training a new directory will be created with the trained models in ./runs/ called train with a number at the end specifying the training iteration
• Run camera_test.py after updating the MODEL_PATH variable at the top with the most recently trained model
• Instead of running the camera you can check validation datasets in the ./runs/train directory showing how validations batches were predicted