This repository contains three versions of Ultralytics YOLO Segmentational models, that trained on CNC-PCB-Grinder dataset for segmentating PCB copper blanks, PCB mashine's workspace and dangerous objects like a clamps, screws, bolts, handlers, special nuts.
best-last.pt - YOLO26n-seg
best-medium - YOLO26m-seg
nano_last_ncnn_model - YOLO26n-seg in NCNN format with graph simplification and FP16 quantization that can be runned in real time on Raspberry Pi 4b with 2 FPS.
Model classes: 0 - workspace 1 - blank 2 - danger
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for CoolFridge/CNC-PCB-Grinder-YOLO-Seg
Base model
Ultralytics/YOLO26