raisinglc

I haven't used nanodet to do detection with onnxruntime. I just used this project to train a model to do classification and usedonnxruntime to run it.

I think there is not much thing which can help you to improve the performance. The training set is very small. It's nearly impossible to achieve the same result without...

把batch size调大 batch size = 1不就是online update，噪声肯定大

Personally, I think the result should be non-negative, so they use torch.abs, although c = sum(y(x)p(x)) where y(x) could be negative. If the output could be negative, it might make...

I add more gpu_id to device setting, change worker number and using torch.DataParallel to map the model to different gpus, This is the simplest way. You could find more detail...

What kind of issue did you encounter? An error or the result of loss is incorrect? Are you training on a single machine with multiple GPUs (torch.DataParallel) or multiple machine...