analysis

Function

Pruning analysis function. It performs only the analysis operations during the pruning process and returns the number of parameters after pruning along with pruning information.

Prototype

analysis(reserved_ratio=0.75, un_prune_list=None)

Parameters

Parameter	Input/Return	Description	Constraints
reserved_ratio	Input	The retention ratio of the number of parameters after pruning.	Optional. Data type: float. Default value: 0.75. Value range: [0, 1].
un_prune_list	Input	Specifies the layers that are not pruned. By default, the first and last layers are not pruned.	Optional. Data type: list, elements must be int or string. Default value: None. If an element is int, it indicates the index of the layer not to be pruned (only Conv2d and Linear operators are counted). If it is string, it indicates the name of the operator in the network.

Sample

import torch
import torchvision
from msmodelslim.pytorch.prune.prune_torch import PruneTorch
model = torchvision.models.vgg16(pretrained=False)
model.eval()
prune_torch = PruneTorch(model, torch.ones([1, 3, 224, 224]).type(torch.float32))
left_params, desc = prune_torch.analysis()
print(desc)
print(left_params)