WongKinYiu / ScaledYOLOv4

Scaled-YOLOv4: Scaling Cross Stage Partial Network
GNU General Public License v3.0
2.01k stars 570 forks source link

resize input size #288

Open namasang1 opened 3 years ago

namasang1 commented 3 years ago

I have a question about the YOLOv4-large.

In the paper, YOLOv4-large models are trained with scaled input image size

so, when i train in these models, should the input image size be large scale (ex 896x896 or 1280x1280) ? Or these models can be resize input image to large input size even if my data size 640x640. because i used custom data which is 640x640 size.

thank you

xcc-1921 commented 3 years ago

I am poor at English. I think the large scale is better than little. BUT it can cause training crash.Because of memory may overflow.

LukeAI commented 3 years ago

there's no substitute for experimenting yourself - try it and see - I think you might see some small improvement with a larger size