search

dwofk / fast-depth

ICRA 2019 "FastDepth: Fast Monocular Depth Estimation on Embedded Systems"
MIT License
926 stars 189 forks source link

Why you saved the entire model instead of state_dict .. ? #59

Open PouyaNF opened 2 years ago

PouyaNF commented 2 years ago

Hi Diana for using torch.load('saved_model') there should be all modules that existed in saving time available otherwise it get an error!!

If you store a model directly via torch.save(model, file_path), you would need to restore the file and folder structure in order to load this model again!!

puyiwen commented 2 years ago

Hi,have you reproduced the author's training code? After I reproduced the project given by the author, the result was not good, and then I did not know why.