I will train custom object detection and segmentation models using yolo


About this gig
Need a model that actually works on your images or video? I train custom computer vision models and deliver clean, reproducible code you can run yourself.
What I can build:
- Object detection, tracking, and counting (YOLO, RT-DETR, Faster R-CNN)
- Instance and semantic segmentation (Mask R-CNN, U-Net, SAM)
- Image classification and defect or quality inspection
- PPE, retail shelf, traffic, and warehouse analytics
- Real-time video pipelines on GPU or edge devices (Jetson, Raspberry Pi)
Every order includes:
- A data check and training plan before I start
- Trained weights plus a ready-to-run inference script
- Metrics (mAP, IoU, precision and recall) with sample predictions
- Clear setup instructions
I've trained models from scratch for paying clients and work daily with PyTorch, so you get a model tuned to your data, not a generic demo.
Stack: Python, PyTorch, Ultralytics, OpenCV, Detectron2, ONNX, TensorRT, FastAPI, Docker.
Please message me with a few sample images or frames before ordering, so I can recommend the right package.
Get to know Daksh A
Machine Learning Engineer
- FromIndia
- Member sinceDec 2020
- Avg. response time1 hour
Languages
English
Other AI Development Services I Offer
FAQ
I don't have labeled data. Can you still help?
Yes. I can advise on labeling tools and formats, or label a small dataset as a custom extra. Message me with the size of your data.
Which annotation formats do you support?
YOLO, COCO, Pascal VOC, and LabelMe. I can convert between them.
Will the model run on my hardware?
Tell me your target (GPU server, CPU, or Jetson or other edge device) and I'll optimize and export for it.
Do I own the model and code?
Yes. You get full source code and trained weights with no lock-in.
