Loading trained model

[adrian70000]

Hi,

I trained my model on my PC and I am trying to load it on esp32 board by using:

model.load_state_dict(torch.load("parameters.pt"))

Which results in:

AttributeError: 'module' object has no attribute 'load'

Am I missing something or is loading pretrained models really not implemented? Where are the model parameters(weights and biases) stored? Can they be exported and imported?

[ljk53]

Hi @adrian70000, right, unfortunately the loading of the pretraind models were not implemented. I might find some time to work on it in the future if people actually find this project useful (it's just a toy project when I started it).