Skip to content

Train and Inference your custom YOLO-NAS model by Single Command Line

Notifications You must be signed in to change notification settings

fsai-dev/fsai-yolo-nas

 
 

Repository files navigation

Setup

  • Create clearml.conf file in base repo dir and add credentials
  • Run make build

Run

  • Run make run

Train

python train.py --name pylon_complex --data /home/data/bboxes/pylon_complex/data.yaml --batch 8 --worker 10 --epoch 300 --model yolo_nas_l --size 1024 --num_gpus 4

Inference

python inference.py --project_name 300_epoch --num_classes 1 --conf 0.5 --source /home/data/bboxes/pylon_complex/val.json --weight /home/yolo-nas-output/runs/pylon_complex/RUN_20240827_232659_396567/ckpt_latest.pth --save --hide

300 epoch run: RUN_20240827_232659_396567

100 epoch run: RUN_20240824_001613_199026

Notes

  • In docker container, you can check super-gradients version by running: python -c 'import super_gradients; print(super_gradients.__version__)'

About

Train and Inference your custom YOLO-NAS model by Single Command Line

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 93.6%
  • Dockerfile 4.7%
  • Makefile 1.7%