Loss and Epocs - Githubissues

iamhankai / ghostnet.pytorch

[CVPR2020] GhostNet: More Features from Cheap Operations

https://arxiv.org/abs/1911.11907

522 stars 116 forks source link

Loss and Epocs #18

Closed dronefreak closed 4 years ago

dronefreak commented 4 years ago

Hi @iamhankai

I am attempting to re-train GhostNet (MobileNetV3) on the ImageNet dataset with some changes in activations and fine-tuning hyperparameters. Could you please tell me how many epochs you trained the network for and what was your final loss value? I could not find them in the paper. Thanks.

iamhankai commented 4 years ago

refer to https://github.com/huawei-noah/ghostnet/issues/9

dronefreak commented 4 years ago

Hi @iamhankai

Thanks for your response. The issue that you have referenced does not mention the information as requested. I am trying to train GhostNet on ImageNet from scratch with modifications for my research. I am very new to Tensorpack usage (and DL in general) so it would be really helpful if I could get the training code for the same.

iamhankai commented 4 years ago

400 epochs. The final loss is 1.9572.

dronefreak commented 4 years ago

Thanks mate.