Checkpoint of the Mobilenet v2 based model

DeepMotionAIResearch / DenseASPP

DenseASPP for Semantic Segmentation in Street Scenes

347 stars 70 forks source link

Open zhixuanli opened 5 years ago

zhixuanli commented 5 years ago

Hi, Are you planning about releasing the checkpoint of the Mobilenet v2 based model?

Without pretrained on ImageNet, I can only get 54.8% on the val set of CityScapes.

Hoping for your pretrained model.

Thank you so much!

mrocr commented 5 years ago

@DeepMotionAIResearch

zhixuanli commented 5 years ago

@DeepMotionAIResearch 😄