Closed tensorboy closed 5 years ago
Did you use the default batchsize?
Yeah, it is 100, right?
maybe I shoud change that default lr, on my 1080Ti, ir_se50 model uses batchsize 100, mobilefacenet uses 200, and you can first find the best batchsize, then use find_lr to locate best lr in your machine, this technique was introduceD by fastai
Great, thanks for your information. I changed it to 200 and retraining now.
closing
Using the default learning rate 0.001 will not work on my side.
Just let you know that I tried 0.01 and that number worked very well. :)