This repo contains the training pipeline for applying style transfer on COCO dataset and finetuning the yolo v8 model on it.
The proposed pipeline for training looks like the following:
The only difference here is I have used the new YOLO V8 model.
To run the entire pipeline on HPC, simply run the following command:
sbatch run_pipeline.sh
This will run the entire pipeline from generating the data to training the model on it.
Note: Ensure that you also have the singularity file in the same folder. You can do that by running this:
rsync -az <yourid>@pioneer.case.edu:/mnt/rds/redhen/gallina/home/rss210/ultralytics_latest.sif .