Torch-Pruning
Torch-Pruning copied to clipboard
How to handle the model with special input?
Our model required more than one input, the forward function of this model could be described as this: ` Class MyModel(nn.Module): def init():
def forward(arg1,arg2,**kwargs):
`
or ` Class MyModel(nn.Module): def init():
def forward(args_dict,**kwargs):
` where kwargs is very important or we need more than one input of the model
How to handle this situation? I tried to put these args to config files and create the model with configs but it seems doesn't work.
Or is there any examples of it?