Closed StereoPancake6 closed 8 months ago
heres some more info: model is put onto RTX 3050 pc (after being trained on gtx 1080ti) and then it is turned from .pt to .engine and then ran.
I've also done every fix recommended, could I have some dependencies missing? improperly installed cuda? ect?
write the command you used to export and train the model
Training: yolo task=detect mode=train imgsz=640 data=CustomCharacterPath.yaml epochs=1000 batch=16 name=Rust
Export: yolo export model="models/rust.pt" format=engine half=true device=0 imgsz=448
heres a little more info, it works fine on my system but once i send it to my friend he has the NMS Issue.
The system it works on: CPU: ryzen 5 5600x GPU: GTX 1080 Ti
The system theres an issue on: CPU: ryzen 3 3200g GPU: RTX 3050
could it be trying to use the ryzen 3 3200g's integrated graphics instead of using the rtx 3050? if so how can i change it?
"WARNING ⚠️ NMS time limit 0.550s exceeded" when on RTX 3050 getting this error makes the ai move the cursor once every 1-2 seconds EXTREMELY slow anyone know a fix? (using a model trained on a gtx 1080 ti)