Open KosukeHao opened 4 years ago
Firstly, thanks for your work!'
When I train yolov4 with my customized data in Linux,
"./darknet detector train build/darknet/x64/data/obj.data cfg/yolo-obj.cfg build/darknet/x64/yolov4.conv.137"
with batch=32, subdivisions=16 , and W\H are all 768, it was ok, but logs shows too quickly.
But, when I tried to train with multi-gpus by command:
./darknet detector train build/darknet/x64/data/obj.data cfg/yolo-obj.cfg build/darknet/x64/yolov4.conv.137 -gpus 0, 1
CUDA Error: out of memory OCCURed.
could anyone tell me the relationship between "batch size" and the "batch" here?
batch=32 mini-batch = batch / subdivisions = 2
when settin: batch=32 subdivisions = 16 mini-batch = batch / subdivisions = 2
1 iterations = 2 or 32 images?
32
Firstly, thanks for your work!'
When I train yolov4 with my customized data in Linux,
with batch=32, subdivisions=16 , and W\H are all 768, it was ok, but logs shows too quickly.
But, when I tried to train with multi-gpus by command:
CUDA Error: out of memory OCCURed.
could anyone tell me the relationship between "batch size" and the "batch" here?