With ‘torch.nn.util.prune’ I can prune torch.nn.conv weight or bias individually, but as I prune a kernel of a conv, the corresponding bias also should be pruned, How could I prune weight and bias collaboratiely of a conv?
With ‘torch.nn.util.prune’ I can prune torch.nn.conv weight or bias individually, but as I prune a kernel of a conv, the corresponding bias also should be pruned, How could I prune weight and bias collaboratiely of a conv?