pretrain model on imagenet
yflv-yanxia opened this issue · 3 comments
yflv-yanxia commented
Thank you for your implementation. Have you released any pre-train model on imagenet database
austingg commented
@yflv-yanxia Currently no pre-trained model. caffe has no fast depthwise conv implemention as tensorflow or mxnet, I strongly recommend Using mxnet, pytorch, or tf to reimplement mobilenetv2. This script is only used for verify the network structure, and maybe in the future, convert a model from other DL Frameworks.
austingg commented
@yflv-yanxia you can reference shicai's mobilenetv2 pretrained model.
yflv-yanxia commented
@austingg Thank you!