Qengineering
Problem when creating a model.bin and model.param
I've tried these two methods to get the ncnn model:
- The python yolov8 method:
from ultralytics import YOLO
import ncnn
model=YOLO('yolov8n.pt')
model.train(data="data.yaml", epochs=100, batch=8)
path = model.export(format="ncnn")
- Converting from onnx to ncnn
from ultralytics import YOLO
model=YOLO('yolov8n.pt')
model.train(data="data.yaml", epochs=100, batch=8)
path = model.export(format="onxx")
After running this code i've done the steps in this guide to obtain ncnn model https://github.com/Tencent/ncnn/wiki/use-ncnn-with-pytorch-or-onnx
None of these methods worked for me and both create a segmentation fault error when running the program.
I'm asking how u have obtained the .bin and .param available in the repo because only them seem to work.
I've tried also to convert not a custom trained model but the standard yolov8nano one but it doesn't work either.
I'll be grateful if someone can explain more in depth how to rearrange the code to work with custom trained model
Did you see this page? https://docs.ultralytics.com/integrations/ncnn/#deploying-exported-yolov8-ncnn-models
Did you see this page? https://docs.ultralytics.com/integrations/ncnn/#deploying-exported-yolov8-ncnn-models
yeah and i did the same as the guide reported
I've used the models from @FeiGeChuanShu see: https://github.com/FeiGeChuanShu/ncnn-android-yolov8