目录

剪枝教程

创建日期:2019年7月22日 | 最后更新日期:2023年11月2日 | 最后验证日期:2024年11月5日

作者: Michela Paganini

最先进的深度学习技术依赖于难以部署的过参数化模型。相反,生物神经网络以使用高效的稀疏连接而闻名。确定通过减少模型中的参数数量来压缩模型的最佳技术对于在不牺牲准确性的情况下减少内存、电池和硬件消耗至关重要。这反过来允许你在设备上部署轻量级模型,并通过设备上的私有计算保证隐私。在研究方面,剪枝被用来研究过参数化网络和欠参数化网络之间学习动态的差异,研究幸运的稀疏子网络和初始化(“彩票票根”)作为破坏性神经架构搜索技术的作用,以及更多。

在本教程中,您将学习如何使用 torch.nn.utils.prune 来稀疏化您的神经网络,并了解如何扩展它以实现您自己的自定义剪枝技术。

需求

"torch>=1.4.0a0+8e8a5e0"

import torch
from torch import nn
import torch.nn.utils.prune as prune
import torch.nn.functional as F

创建模型

在本教程中,我们使用了LeCun等人于1998年提出的LeNet架构。

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

class LeNet(nn.Module):
    def __init__(self):
        super(LeNet, self).__init__()
        # 1 input image channel, 6 output channels, 5x5 square conv kernel
        self.conv1 = nn.Conv2d(1, 6, 5)
        self.conv2 = nn.Conv2d(6, 16, 5)
        self.fc1 = nn.Linear(16 * 5 * 5, 120)  # 5x5 image dimension
        self.fc2 = nn.Linear(120, 84)
        self.fc3 = nn.Linear(84, 10)

    def forward(self, x):
        x = F.max_pool2d(F.relu(self.conv1(x)), (2, 2))
        x = F.max_pool2d(F.relu(self.conv2(x)), 2)
        x = x.view(-1, int(x.nelement() / x.shape[0]))
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        x = self.fc3(x)
        return x

model = LeNet().to(device=device)

检查一个模块

让我们检查一下(未剪枝的)conv1层在我们的LeNet模型中。它将包含两个 参数 weightbias,目前没有缓冲区。

[('weight', Parameter containing:
tensor([[[[ 0.1529,  0.1660, -0.0469,  0.1837, -0.0438],
          [ 0.0404, -0.0974,  0.1175,  0.1763, -0.1467],
          [ 0.1738,  0.0374,  0.1478,  0.0271,  0.0964],
          [-0.0282,  0.1542,  0.0296, -0.0934,  0.0510],
          [-0.0921, -0.0235, -0.0812,  0.1327, -0.1579]]],


        [[[-0.0922, -0.0565, -0.1203,  0.0189, -0.1975],
          [ 0.1806, -0.1699,  0.1544,  0.0333, -0.0649],
          [ 0.1236,  0.0312,  0.1616,  0.0219, -0.0631],
          [ 0.0537, -0.0542,  0.0842,  0.1786,  0.1156],
          [-0.0874,  0.1155,  0.0358,  0.1016, -0.1219]]],


        [[[-0.1980, -0.0773, -0.1534,  0.1641,  0.0576],
          [ 0.0828,  0.0633, -0.0035,  0.1565, -0.1421],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0667,  0.1925, -0.1651, -0.1984]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.1033, -0.1363,  0.1061, -0.0808,  0.1214],
          [-0.0475,  0.1144, -0.1554, -0.1009,  0.0610],
          [ 0.0423, -0.0510,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0675, -0.0709, -0.1935]]],


        [[[-0.1145,  0.0500, -0.0264, -0.1452,  0.0047],
          [-0.1366, -0.1697, -0.1101, -0.1750, -0.1273],
          [ 0.1999,  0.0378,  0.0616, -0.1865, -0.1314],
          [-0.0666,  0.0313, -0.1760, -0.0862, -0.1197],
          [ 0.0006, -0.0744, -0.0139, -0.1355, -0.1373]]],


        [[[-0.1167, -0.0685, -0.1579,  0.1677, -0.0397],
          [ 0.1721,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.1988,  0.0572, -0.0437],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.1033,  0.1615,  0.1822, -0.1586]]]], device='cuda:0',
       requires_grad=True)), ('bias', Parameter containing:
tensor([ 0.0503, -0.0860, -0.0219, -0.1497,  0.1822, -0.1468], device='cuda:0',
       requires_grad=True))]
print(list(module.named_buffers()))
[]

剪枝一个模块

要修剪一个模块(在此示例中,是我们 LeNet 架构的第 conv1 层),首先从 torch.nn.utils.prune 中选择一种修剪技术(或者通过继承 BasePruningMethod 自行实现)。然后,指定该模块及其内部需要修剪的参数名称。最后,使用所选修剪技术所需的适当关键字参数,指定修剪参数。

在这个示例中,我们将随机剪枝 weight 层中名为 conv1 的参数 30% 的连接。 该模块作为函数的第一个参数传递;name 使用其字符串标识符在该模块内识别参数;而 amount 表示要剪枝的连接百分比(如果是一个介于 0. 和 1. 之间的浮点数),或要剪枝的连接绝对数量(如果是一个非负整数)。

prune.random_unstructured(module, name="weight", amount=0.3)
Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1))

剪枝通过从参数中移除 weight 并将其替换为 一个名为 weight_orig 的新参数(即在初始参数 name 后追加 "_orig")。weight_orig 存储了未剪枝的张量版本。由于 bias 没有被剪枝,因此它将保持完整。

print(list(module.named_parameters()))
[('bias', Parameter containing:
tensor([ 0.0503, -0.0860, -0.0219, -0.1497,  0.1822, -0.1468], device='cuda:0',
       requires_grad=True)), ('weight_orig', Parameter containing:
tensor([[[[ 0.1529,  0.1660, -0.0469,  0.1837, -0.0438],
          [ 0.0404, -0.0974,  0.1175,  0.1763, -0.1467],
          [ 0.1738,  0.0374,  0.1478,  0.0271,  0.0964],
          [-0.0282,  0.1542,  0.0296, -0.0934,  0.0510],
          [-0.0921, -0.0235, -0.0812,  0.1327, -0.1579]]],


        [[[-0.0922, -0.0565, -0.1203,  0.0189, -0.1975],
          [ 0.1806, -0.1699,  0.1544,  0.0333, -0.0649],
          [ 0.1236,  0.0312,  0.1616,  0.0219, -0.0631],
          [ 0.0537, -0.0542,  0.0842,  0.1786,  0.1156],
          [-0.0874,  0.1155,  0.0358,  0.1016, -0.1219]]],


        [[[-0.1980, -0.0773, -0.1534,  0.1641,  0.0576],
          [ 0.0828,  0.0633, -0.0035,  0.1565, -0.1421],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0667,  0.1925, -0.1651, -0.1984]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.1033, -0.1363,  0.1061, -0.0808,  0.1214],
          [-0.0475,  0.1144, -0.1554, -0.1009,  0.0610],
          [ 0.0423, -0.0510,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0675, -0.0709, -0.1935]]],


        [[[-0.1145,  0.0500, -0.0264, -0.1452,  0.0047],
          [-0.1366, -0.1697, -0.1101, -0.1750, -0.1273],
          [ 0.1999,  0.0378,  0.0616, -0.1865, -0.1314],
          [-0.0666,  0.0313, -0.1760, -0.0862, -0.1197],
          [ 0.0006, -0.0744, -0.0139, -0.1355, -0.1373]]],


        [[[-0.1167, -0.0685, -0.1579,  0.1677, -0.0397],
          [ 0.1721,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.1988,  0.0572, -0.0437],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.1033,  0.1615,  0.1822, -0.1586]]]], device='cuda:0',
       requires_grad=True))]

剪枝技术生成的剪枝掩码以上述选择的技术保存 为一个名为weight_mask的模块缓冲区(即在初始参数name后附加"_mask")。

print(list(module.named_buffers()))
[('weight_mask', tensor([[[[1., 1., 1., 1., 1.],
          [1., 0., 1., 1., 1.],
          [1., 0., 0., 1., 1.],
          [1., 0., 1., 1., 1.],
          [1., 0., 0., 1., 1.]]],


        [[[1., 1., 1., 0., 1.],
          [1., 1., 1., 1., 1.],
          [0., 1., 1., 1., 0.],
          [1., 1., 0., 1., 0.],
          [0., 1., 0., 1., 1.]]],


        [[[1., 0., 0., 0., 1.],
          [1., 0., 1., 1., 0.],
          [1., 1., 1., 1., 1.],
          [1., 1., 1., 1., 1.],
          [1., 0., 1., 1., 0.]]],


        [[[1., 1., 1., 1., 1.],
          [0., 1., 1., 1., 0.],
          [1., 1., 1., 0., 1.],
          [0., 0., 1., 1., 1.],
          [1., 1., 0., 1., 1.]]],


        [[[1., 0., 1., 1., 1.],
          [1., 1., 0., 0., 0.],
          [1., 1., 0., 0., 0.],
          [0., 1., 1., 0., 1.],
          [1., 0., 0., 0., 1.]]],


        [[[1., 0., 1., 0., 1.],
          [0., 1., 1., 1., 1.],
          [1., 1., 0., 1., 0.],
          [1., 1., 1., 1., 1.],
          [1., 0., 0., 1., 1.]]]], device='cuda:0'))]

为了正向传播无需修改,weight 属性 需要存在。在torch.nn.utils.prune 中实现的剪枝技术计算权重的剪枝版本(通过 将掩码与原始参数结合)并存储在属性 weight 中。注意,这不再是 module 的参数,现在只是一个属性。

tensor([[[[ 0.1529,  0.1660, -0.0469,  0.1837, -0.0438],
          [ 0.0404, -0.0000,  0.1175,  0.1763, -0.1467],
          [ 0.1738,  0.0000,  0.0000,  0.0271,  0.0964],
          [-0.0282,  0.0000,  0.0296, -0.0934,  0.0510],
          [-0.0921, -0.0000, -0.0000,  0.1327, -0.1579]]],


        [[[-0.0922, -0.0565, -0.1203,  0.0000, -0.1975],
          [ 0.1806, -0.1699,  0.1544,  0.0333, -0.0649],
          [ 0.0000,  0.0312,  0.1616,  0.0219, -0.0000],
          [ 0.0537, -0.0542,  0.0000,  0.1786,  0.0000],
          [-0.0000,  0.1155,  0.0000,  0.1016, -0.1219]]],


        [[[-0.1980, -0.0000, -0.0000,  0.0000,  0.0576],
          [ 0.0828,  0.0000, -0.0035,  0.1565, -0.0000],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0000,  0.1925, -0.1651, -0.0000]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.0000, -0.1363,  0.1061, -0.0808,  0.0000],
          [-0.0475,  0.1144, -0.1554, -0.0000,  0.0610],
          [ 0.0000, -0.0000,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0000, -0.0709, -0.1935]]],


        [[[-0.1145,  0.0000, -0.0264, -0.1452,  0.0047],
          [-0.1366, -0.1697, -0.0000, -0.0000, -0.0000],
          [ 0.1999,  0.0378,  0.0000, -0.0000, -0.0000],
          [-0.0000,  0.0313, -0.1760, -0.0000, -0.1197],
          [ 0.0006, -0.0000, -0.0000, -0.0000, -0.1373]]],


        [[[-0.1167, -0.0000, -0.1579,  0.0000, -0.0397],
          [ 0.0000,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.0000,  0.0572, -0.0000],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.0000,  0.0000,  0.1822, -0.1586]]]], device='cuda:0',
       grad_fn=<MulBackward0>)

最后,在每次前向传递之前都会应用剪枝,使用PyTorch的 forward_pre_hooks。具体来说,当module被剪枝时,如我们在这里所做的,它将为每个被剪枝的参数获取一个forward_pre_hook。在这种情况下,由于到目前为止我们只剪枝了原始参数weight,因此只会存在一个钩子。

print(module._forward_pre_hooks)
OrderedDict([(0, <torch.nn.utils.prune.RandomUnstructured object at 0x7fce471ea0b0>)])

为了完整性,我们现在也可以修剪 bias,以查看 module 的参数、缓冲区、钩子和属性如何变化。 仅仅为了尝试另一种修剪技术,这里我们按L1范数修剪偏置中最小的3个条目,如 l1_unstructured 修剪函数中实现的那样。

prune.l1_unstructured(module, name="bias", amount=3)
Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1))

我们现在期望命名参数中包含 weight_orig(来自之前)和 bias_orig。缓冲区将包括 weight_maskbias_mask。两个张量的剪枝版本将作为模块属性存在,模块现在将有两个 forward_pre_hooks

print(list(module.named_parameters()))
[('weight_orig', Parameter containing:
tensor([[[[ 0.1529,  0.1660, -0.0469,  0.1837, -0.0438],
          [ 0.0404, -0.0974,  0.1175,  0.1763, -0.1467],
          [ 0.1738,  0.0374,  0.1478,  0.0271,  0.0964],
          [-0.0282,  0.1542,  0.0296, -0.0934,  0.0510],
          [-0.0921, -0.0235, -0.0812,  0.1327, -0.1579]]],


        [[[-0.0922, -0.0565, -0.1203,  0.0189, -0.1975],
          [ 0.1806, -0.1699,  0.1544,  0.0333, -0.0649],
          [ 0.1236,  0.0312,  0.1616,  0.0219, -0.0631],
          [ 0.0537, -0.0542,  0.0842,  0.1786,  0.1156],
          [-0.0874,  0.1155,  0.0358,  0.1016, -0.1219]]],


        [[[-0.1980, -0.0773, -0.1534,  0.1641,  0.0576],
          [ 0.0828,  0.0633, -0.0035,  0.1565, -0.1421],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0667,  0.1925, -0.1651, -0.1984]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.1033, -0.1363,  0.1061, -0.0808,  0.1214],
          [-0.0475,  0.1144, -0.1554, -0.1009,  0.0610],
          [ 0.0423, -0.0510,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0675, -0.0709, -0.1935]]],


        [[[-0.1145,  0.0500, -0.0264, -0.1452,  0.0047],
          [-0.1366, -0.1697, -0.1101, -0.1750, -0.1273],
          [ 0.1999,  0.0378,  0.0616, -0.1865, -0.1314],
          [-0.0666,  0.0313, -0.1760, -0.0862, -0.1197],
          [ 0.0006, -0.0744, -0.0139, -0.1355, -0.1373]]],


        [[[-0.1167, -0.0685, -0.1579,  0.1677, -0.0397],
          [ 0.1721,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.1988,  0.0572, -0.0437],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.1033,  0.1615,  0.1822, -0.1586]]]], device='cuda:0',
       requires_grad=True)), ('bias_orig', Parameter containing:
tensor([ 0.0503, -0.0860, -0.0219, -0.1497,  0.1822, -0.1468], device='cuda:0',
       requires_grad=True))]
print(list(module.named_buffers()))
[('weight_mask', tensor([[[[1., 1., 1., 1., 1.],
          [1., 0., 1., 1., 1.],
          [1., 0., 0., 1., 1.],
          [1., 0., 1., 1., 1.],
          [1., 0., 0., 1., 1.]]],


        [[[1., 1., 1., 0., 1.],
          [1., 1., 1., 1., 1.],
          [0., 1., 1., 1., 0.],
          [1., 1., 0., 1., 0.],
          [0., 1., 0., 1., 1.]]],


        [[[1., 0., 0., 0., 1.],
          [1., 0., 1., 1., 0.],
          [1., 1., 1., 1., 1.],
          [1., 1., 1., 1., 1.],
          [1., 0., 1., 1., 0.]]],


        [[[1., 1., 1., 1., 1.],
          [0., 1., 1., 1., 0.],
          [1., 1., 1., 0., 1.],
          [0., 0., 1., 1., 1.],
          [1., 1., 0., 1., 1.]]],


        [[[1., 0., 1., 1., 1.],
          [1., 1., 0., 0., 0.],
          [1., 1., 0., 0., 0.],
          [0., 1., 1., 0., 1.],
          [1., 0., 0., 0., 1.]]],


        [[[1., 0., 1., 0., 1.],
          [0., 1., 1., 1., 1.],
          [1., 1., 0., 1., 0.],
          [1., 1., 1., 1., 1.],
          [1., 0., 0., 1., 1.]]]], device='cuda:0')), ('bias_mask', tensor([0., 0., 0., 1., 1., 1.], device='cuda:0'))]
print(module.bias)
tensor([ 0.0000, -0.0000, -0.0000, -0.1497,  0.1822, -0.1468], device='cuda:0',
       grad_fn=<MulBackward0>)
print(module._forward_pre_hooks)
OrderedDict([(0, <torch.nn.utils.prune.RandomUnstructured object at 0x7fce471ea0b0>), (1, <torch.nn.utils.prune.L1Unstructured object at 0x7fce471e98a0>)])

迭代剪枝

同一个模块中的参数可以被多次剪枝,各种剪枝调用的效果等同于依次应用的各种掩码的组合。 新的掩码与旧的掩码的组合由PruningContainer’s compute_mask方法处理。

例如,假设我们现在想进一步剪枝 module.weight,这次使用沿张量0轴的结构化剪枝(0轴对应卷积层的输出通道,维度为 conv1),基于通道的L2范数。这可以通过使用 ln_structured 函数,结合 n=2dim=0 来实现。

prune.ln_structured(module, name="weight", amount=0.5, n=2, dim=0)

# As we can verify, this will zero out all the connections corresponding to
# 50% (3 out of 6) of the channels, while preserving the action of the
# previous mask.
print(module.weight)
tensor([[[[ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000],
          [-0.0000,  0.0000,  0.0000, -0.0000,  0.0000],
          [-0.0000, -0.0000, -0.0000,  0.0000, -0.0000]]],


        [[[-0.0000, -0.0000, -0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000,  0.0000],
          [-0.0000,  0.0000,  0.0000,  0.0000, -0.0000]]],


        [[[-0.1980, -0.0000, -0.0000,  0.0000,  0.0576],
          [ 0.0828,  0.0000, -0.0035,  0.1565, -0.0000],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0000,  0.1925, -0.1651, -0.0000]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.0000, -0.1363,  0.1061, -0.0808,  0.0000],
          [-0.0475,  0.1144, -0.1554, -0.0000,  0.0610],
          [ 0.0000, -0.0000,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0000, -0.0709, -0.1935]]],


        [[[-0.0000,  0.0000, -0.0000, -0.0000,  0.0000],
          [-0.0000, -0.0000, -0.0000, -0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000, -0.0000, -0.0000],
          [-0.0000,  0.0000, -0.0000, -0.0000, -0.0000],
          [ 0.0000, -0.0000, -0.0000, -0.0000, -0.0000]]],


        [[[-0.1167, -0.0000, -0.1579,  0.0000, -0.0397],
          [ 0.0000,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.0000,  0.0572, -0.0000],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.0000,  0.0000,  0.1822, -0.1586]]]], device='cuda:0',
       grad_fn=<MulBackward0>)

对应的钩子现在将是类型 torch.nn.utils.prune.PruningContainer,并将存储对 weight 参数应用的剪枝历史。

for hook in module._forward_pre_hooks.values():
    if hook._tensor_name == "weight":  # select out the correct hook
        break

print(list(hook))  # pruning history in the container
[<torch.nn.utils.prune.RandomUnstructured object at 0x7fce471ea0b0>, <torch.nn.utils.prune.LnStructured object at 0x7fce471e9a20>]

对剪枝模型进行序列化

所有相关的张量,包括掩码缓冲区和用于计算剪枝张量的原始参数 都存储在模型的 state_dict 因此,如果需要,可以轻松地进行序列化和保存。

print(model.state_dict().keys())
odict_keys(['conv1.weight_orig', 'conv1.bias_orig', 'conv1.weight_mask', 'conv1.bias_mask', 'conv2.weight', 'conv2.bias', 'fc1.weight', 'fc1.bias', 'fc2.weight', 'fc2.bias', 'fc3.weight', 'fc3.bias'])

移除剪枝重参数化

为了使剪枝永久生效,请移除第weight_origweight_mask中的重参数化,并移除forward_pre_hook,我们可以使用remove功能来自torch.nn.utils.prune。请注意,这不会撤销剪枝,因为就像它从未发生过一样。它只是通过将参数weight重新分配给其剪枝后的模型参数,使其成为永久性的。

在移除重参数化之前:

print(list(module.named_parameters()))
[('weight_orig', Parameter containing:
tensor([[[[ 0.1529,  0.1660, -0.0469,  0.1837, -0.0438],
          [ 0.0404, -0.0974,  0.1175,  0.1763, -0.1467],
          [ 0.1738,  0.0374,  0.1478,  0.0271,  0.0964],
          [-0.0282,  0.1542,  0.0296, -0.0934,  0.0510],
          [-0.0921, -0.0235, -0.0812,  0.1327, -0.1579]]],


        [[[-0.0922, -0.0565, -0.1203,  0.0189, -0.1975],
          [ 0.1806, -0.1699,  0.1544,  0.0333, -0.0649],
          [ 0.1236,  0.0312,  0.1616,  0.0219, -0.0631],
          [ 0.0537, -0.0542,  0.0842,  0.1786,  0.1156],
          [-0.0874,  0.1155,  0.0358,  0.1016, -0.1219]]],


        [[[-0.1980, -0.0773, -0.1534,  0.1641,  0.0576],
          [ 0.0828,  0.0633, -0.0035,  0.1565, -0.1421],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0667,  0.1925, -0.1651, -0.1984]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.1033, -0.1363,  0.1061, -0.0808,  0.1214],
          [-0.0475,  0.1144, -0.1554, -0.1009,  0.0610],
          [ 0.0423, -0.0510,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0675, -0.0709, -0.1935]]],


        [[[-0.1145,  0.0500, -0.0264, -0.1452,  0.0047],
          [-0.1366, -0.1697, -0.1101, -0.1750, -0.1273],
          [ 0.1999,  0.0378,  0.0616, -0.1865, -0.1314],
          [-0.0666,  0.0313, -0.1760, -0.0862, -0.1197],
          [ 0.0006, -0.0744, -0.0139, -0.1355, -0.1373]]],


        [[[-0.1167, -0.0685, -0.1579,  0.1677, -0.0397],
          [ 0.1721,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.1988,  0.0572, -0.0437],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.1033,  0.1615,  0.1822, -0.1586]]]], device='cuda:0',
       requires_grad=True)), ('bias_orig', Parameter containing:
tensor([ 0.0503, -0.0860, -0.0219, -0.1497,  0.1822, -0.1468], device='cuda:0',
       requires_grad=True))]
print(list(module.named_buffers()))
[('weight_mask', tensor([[[[0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.]]],


        [[[0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.]]],


        [[[1., 0., 0., 0., 1.],
          [1., 0., 1., 1., 0.],
          [1., 1., 1., 1., 1.],
          [1., 1., 1., 1., 1.],
          [1., 0., 1., 1., 0.]]],


        [[[1., 1., 1., 1., 1.],
          [0., 1., 1., 1., 0.],
          [1., 1., 1., 0., 1.],
          [0., 0., 1., 1., 1.],
          [1., 1., 0., 1., 1.]]],


        [[[0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.],
          [0., 0., 0., 0., 0.]]],


        [[[1., 0., 1., 0., 1.],
          [0., 1., 1., 1., 1.],
          [1., 1., 0., 1., 0.],
          [1., 1., 1., 1., 1.],
          [1., 0., 0., 1., 1.]]]], device='cuda:0')), ('bias_mask', tensor([0., 0., 0., 1., 1., 1.], device='cuda:0'))]
tensor([[[[ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000],
          [-0.0000,  0.0000,  0.0000, -0.0000,  0.0000],
          [-0.0000, -0.0000, -0.0000,  0.0000, -0.0000]]],


        [[[-0.0000, -0.0000, -0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000,  0.0000],
          [-0.0000,  0.0000,  0.0000,  0.0000, -0.0000]]],


        [[[-0.1980, -0.0000, -0.0000,  0.0000,  0.0576],
          [ 0.0828,  0.0000, -0.0035,  0.1565, -0.0000],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0000,  0.1925, -0.1651, -0.0000]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.0000, -0.1363,  0.1061, -0.0808,  0.0000],
          [-0.0475,  0.1144, -0.1554, -0.0000,  0.0610],
          [ 0.0000, -0.0000,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0000, -0.0709, -0.1935]]],


        [[[-0.0000,  0.0000, -0.0000, -0.0000,  0.0000],
          [-0.0000, -0.0000, -0.0000, -0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000, -0.0000, -0.0000],
          [-0.0000,  0.0000, -0.0000, -0.0000, -0.0000],
          [ 0.0000, -0.0000, -0.0000, -0.0000, -0.0000]]],


        [[[-0.1167, -0.0000, -0.1579,  0.0000, -0.0397],
          [ 0.0000,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.0000,  0.0572, -0.0000],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.0000,  0.0000,  0.1822, -0.1586]]]], device='cuda:0',
       grad_fn=<MulBackward0>)

在移除重参数化之后:

prune.remove(module, 'weight')
print(list(module.named_parameters()))
[('bias_orig', Parameter containing:
tensor([ 0.0503, -0.0860, -0.0219, -0.1497,  0.1822, -0.1468], device='cuda:0',
       requires_grad=True)), ('weight', Parameter containing:
tensor([[[[ 0.0000,  0.0000, -0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000],
          [-0.0000,  0.0000,  0.0000, -0.0000,  0.0000],
          [-0.0000, -0.0000, -0.0000,  0.0000, -0.0000]]],


        [[[-0.0000, -0.0000, -0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000,  0.0000, -0.0000],
          [ 0.0000, -0.0000,  0.0000,  0.0000,  0.0000],
          [-0.0000,  0.0000,  0.0000,  0.0000, -0.0000]]],


        [[[-0.1980, -0.0000, -0.0000,  0.0000,  0.0576],
          [ 0.0828,  0.0000, -0.0035,  0.1565, -0.0000],
          [ 0.0126, -0.1365,  0.0617, -0.0689,  0.0613],
          [-0.0417,  0.1659, -0.1185, -0.1193, -0.1193],
          [ 0.1799,  0.0000,  0.1925, -0.1651, -0.0000]]],


        [[[-0.1565, -0.1345,  0.0810,  0.0716,  0.1662],
          [-0.0000, -0.1363,  0.1061, -0.0808,  0.0000],
          [-0.0475,  0.1144, -0.1554, -0.0000,  0.0610],
          [ 0.0000, -0.0000,  0.1192,  0.1360, -0.1450],
          [-0.1068,  0.1831, -0.0000, -0.0709, -0.1935]]],


        [[[-0.0000,  0.0000, -0.0000, -0.0000,  0.0000],
          [-0.0000, -0.0000, -0.0000, -0.0000, -0.0000],
          [ 0.0000,  0.0000,  0.0000, -0.0000, -0.0000],
          [-0.0000,  0.0000, -0.0000, -0.0000, -0.0000],
          [ 0.0000, -0.0000, -0.0000, -0.0000, -0.0000]]],


        [[[-0.1167, -0.0000, -0.1579,  0.0000, -0.0397],
          [ 0.0000,  0.0623, -0.1694,  0.1384, -0.0550],
          [-0.0767, -0.1660, -0.0000,  0.0572, -0.0000],
          [ 0.0779, -0.1641,  0.1485, -0.1468, -0.0345],
          [ 0.0418,  0.0000,  0.0000,  0.1822, -0.1586]]]], device='cuda:0',
       requires_grad=True))]
print(list(module.named_buffers()))
[('bias_mask', tensor([0., 0., 0., 1., 1., 1.], device='cuda:0'))]

模型中多个参数的剪枝

通过指定所需的剪枝技术及参数,我们可以轻松地对网络中的多个张量进行剪枝,也许根据它们的类型,如本例中将要看到的那样。

new_model = LeNet()
for name, module in new_model.named_modules():
    # prune 20% of connections in all 2D-conv layers
    if isinstance(module, torch.nn.Conv2d):
        prune.l1_unstructured(module, name='weight', amount=0.2)
    # prune 40% of connections in all linear layers
    elif isinstance(module, torch.nn.Linear):
        prune.l1_unstructured(module, name='weight', amount=0.4)

print(dict(new_model.named_buffers()).keys())  # to verify that all masks exist
dict_keys(['conv1.weight_mask', 'conv2.weight_mask', 'fc1.weight_mask', 'fc2.weight_mask', 'fc3.weight_mask'])

全局剪枝

到目前为止,我们只关注了通常被称为“局部”剪枝的内容, 即通过将每个张量中的条目与其他条目进行比较(如权重幅度、激活值、梯度等),逐个剪枝张量。然而,一种常见且可能更强大的技术是将整个模型一次性剪枝,例如移除整个模型中最低的20%的连接,而不是在每一层中分别移除最低的20%的连接。这可能导致各层不同的剪枝比例。让我们看看如何使用 global_unstructuredtorch.nn.utils.prune 来实现这一点。

model = LeNet()

parameters_to_prune = (
    (model.conv1, 'weight'),
    (model.conv2, 'weight'),
    (model.fc1, 'weight'),
    (model.fc2, 'weight'),
    (model.fc3, 'weight'),
)

prune.global_unstructured(
    parameters_to_prune,
    pruning_method=prune.L1Unstructured,
    amount=0.2,
)

现在我们可以检查每个剪枝参数所引入的稀疏性,这在每一层不会等于 20%。然而,全局稀疏性将大约为 20%。

print(
    "Sparsity in conv1.weight: {:.2f}%".format(
        100. * float(torch.sum(model.conv1.weight == 0))
        / float(model.conv1.weight.nelement())
    )
)
print(
    "Sparsity in conv2.weight: {:.2f}%".format(
        100. * float(torch.sum(model.conv2.weight == 0))
        / float(model.conv2.weight.nelement())
    )
)
print(
    "Sparsity in fc1.weight: {:.2f}%".format(
        100. * float(torch.sum(model.fc1.weight == 0))
        / float(model.fc1.weight.nelement())
    )
)
print(
    "Sparsity in fc2.weight: {:.2f}%".format(
        100. * float(torch.sum(model.fc2.weight == 0))
        / float(model.fc2.weight.nelement())
    )
)
print(
    "Sparsity in fc3.weight: {:.2f}%".format(
        100. * float(torch.sum(model.fc3.weight == 0))
        / float(model.fc3.weight.nelement())
    )
)
print(
    "Global sparsity: {:.2f}%".format(
        100. * float(
            torch.sum(model.conv1.weight == 0)
            + torch.sum(model.conv2.weight == 0)
            + torch.sum(model.fc1.weight == 0)
            + torch.sum(model.fc2.weight == 0)
            + torch.sum(model.fc3.weight == 0)
        )
        / float(
            model.conv1.weight.nelement()
            + model.conv2.weight.nelement()
            + model.fc1.weight.nelement()
            + model.fc2.weight.nelement()
            + model.fc3.weight.nelement()
        )
    )
)
Sparsity in conv1.weight: 4.67%
Sparsity in conv2.weight: 13.92%
Sparsity in fc1.weight: 22.16%
Sparsity in fc2.weight: 12.10%
Sparsity in fc3.weight: 11.31%
Global sparsity: 20.00%

扩展 torch.nn.utils.prune 使用自定义剪枝函数

要实现自己的剪枝函数,您可以继承 nn.utils.prune 模块并通过子类化 BasePruningMethod 基类来实现,与其他所有剪枝方法的方式相同。基类 为您实现了以下方法:__call__apply_maskapplypruneremove。除了某些特殊情况外,您不需要 为新的剪枝技术重新实现这些方法。 您将需要实现 __init__(构造函数), 以及 compute_mask(根据您的剪枝 技术逻辑计算给定张量的掩码的说明)。此外,您还需要指定该技术实现的是哪种 剪枝类型(支持的选项包括 globalstructuredunstructured)。这是为了确定 在迭代应用剪枝时如何组合掩码。换句话说,当剪枝一个已经剪枝过的参数时, 当前剪枝技术应作用于该参数未剪枝的部分。指定 PRUNING_TYPE 将 使 PruningContainer(处理剪枝掩码的迭代应用)能够正确识别 要剪枝的参数切片。

假设,例如,你想实现一种剪枝技术,该技术会剪枝张量中的每隔一个元素(或者——如果张量之前已经被剪枝过——则在剩余未剪枝的部分中每隔一个元素)。这将是PRUNING_TYPE='unstructured',因为它作用于层中的单个连接,而不是整个单元/通道('structured'),或跨不同参数('global')。

class FooBarPruningMethod(prune.BasePruningMethod):
    """Prune every other entry in a tensor
    """
    PRUNING_TYPE = 'unstructured'

    def compute_mask(self, t, default_mask):
        mask = default_mask.clone()
        mask.view(-1)[::2] = 0
        return mask

现在,要将此应用到 nn.Module 中的一个参数,您还应 提供一个简单的函数来实例化该方法并 应用它。

def foobar_unstructured(module, name):
    """Prunes tensor corresponding to parameter called `name` in `module`
    by removing every other entry in the tensors.
    Modifies module in place (and also return the modified module)
    by:
    1) adding a named buffer called `name+'_mask'` corresponding to the
    binary mask applied to the parameter `name` by the pruning method.
    The parameter `name` is replaced by its pruned version, while the
    original (unpruned) parameter is stored in a new parameter named
    `name+'_orig'`.

    Args:
        module (nn.Module): module containing the tensor to prune
        name (string): parameter name within `module` on which pruning
                will act.

    Returns:
        module (nn.Module): modified (i.e. pruned) version of the input
            module

    Examples:
        >>> m = nn.Linear(3, 4)
        >>> foobar_unstructured(m, name='bias')
    """
    FooBarPruningMethod.apply(module, name)
    return module

让我们来试试看!

model = LeNet()
foobar_unstructured(model.fc3, name='bias')

print(model.fc3.bias_mask)
tensor([0., 1., 0., 1., 0., 1., 0., 1., 0., 1.])

脚本总运行时间: ( 0 分钟 0.281 秒)

通过 Sphinx-Gallery 生成的画廊

文档

访问 PyTorch 的全面开发人员文档

查看文档

教程

获取面向初学者和高级开发人员的深入教程

查看教程

资源

查找开发资源并解答您的问题

查看资源