Saving and loading a model in Pytorch?

[pinocchio (Rene Sandoval)]

What is wrong with doing:

def save_ckpt(path_to_ckpt):
    from pathlib import Path
    import dill as pickle
    ## Make dir. Throw no exceptions if it already exists
    path_to_ckpt.mkdir(parents=True, exist_ok=True)
    ckpt_path_plus_path = path_to_ckpt / Path('db')

    ## Pickle args
    db['crazy_mdl'] = crazy_mdl
    with open(ckpt_path_plus_path , 'ab') as db_file:
        pickle.dump(db, db_file)

How does one pickle arbitrary pytorch models that use lambda functions?

I currently have a neural network module: import torch.nn as nn class NN(nn.Module): def __init__(self,args,lambda_f,nn1, loss, opt): super().__init__() self.args = args self.lambda_f = lambda_f self.nn1 = nn1 self.loss = loss self.opt = opt # more nn.Params stuff etc... def forward(self, x): #some code using fields return out I am trying to checkpoint it but because pytorch saves using state_dicts it means I can…

---

Advantages & Disadvantages of using pickle module to save models vs torch.save

I was reading Save and Load model but it wasn’t clear why I’d use torch.save over pickle.dump. What worries me is that my Neural Net modules/object have many more things inside of them besides only parameters. Since pytorch saves things using the state_dict that worried me that something might be missing (https://pytorch.org/tutorials/beginner/saving_loading_models.html). Isn’t for my use case easier to simply use the pickle.dump function? Why would I use torch.save in general? Note I am usin…

---

Save and Load model

what are the advantages and disadvantages of just using the python pickle @ptrblck what are the advantages and disadvantages of just using the python pickle module vs the load/save stuff pytorch has built?