模型优化点

1. SWA(随机权重平均)

简单来说就是将指定epochs范围内训练出来的模型，将这些模型的weight取平均赋值给这个swa_model。

1. Pytorch内置的swa，但是引入了一个SWALR。
2. 一个项目用到的。

2. attack training(对抗训练)

import torch
import torch.nn as nn


# FGM
class FGM:
    def __init__(self, model: nn.Module, eps=1.):
        self.model = (
            model.module if hasattr(model, "module") else model
        )
        self.eps = eps
        self.backup = {}

    # only attack word embedding
    def attack(self, emb_name='word_embeddings'):
        for name, param in self.model.named_parameters():
            if param.requires_grad and emb_name in name:
                self.backup[name] = param.data.clone()
                norm = torch.norm(param.grad)
                if norm and not torch.isnan(norm):
                    r_at = self.eps * param.grad / norm
                    param.data.add_(r_at)

    def restore(self, emb_name='word_embeddings'):
        for name, para in self.model.named_parameters():
            if para.requires_grad and emb_name in name:
                assert name in self.backup
                para.data = self.backup[name]

        self.backup = {}


# PGD
class PGD:
    def __init__(self, model, eps=1., alpha=0.3):
        self.model = (
            model.module if hasattr(model, "module") else model
        )
        self.eps = eps
        self.alpha = alpha
        self.emb_backup = {}
        self.grad_backup = {}

    def attack(self, emb_name='word_embeddings', is_first_attack=False):
        for name, param in self.model.named_parameters():
            if param.requires_grad and emb_name in name:
                if is_first_attack:
                    self.emb_backup[name] = param.data.clone()
                norm = torch.norm(param.grad)
                if norm != 0 and not torch.isnan(norm):
                    r_at = self.alpha * param.grad / norm
                    param.data.add_(r_at)
                    param.data = self.project(name, param.data)

    def restore(self, emb_name='word_embeddings'):
        for name, param in self.model.named_parameters():
            if param.requires_grad and emb_name in name:
                assert name in self.emb_backup
                param.data = self.emb_backup[name]
        self.emb_backup = {}

    def project(self, param_name, param_data):
        r = param_data - self.emb_backup[param_name]
        if torch.norm(r) > self.eps:
            r = self.eps * r / torch.norm(r)
        return self.emb_backup[param_name] + r

    def backup_grad(self):
        for name, param in self.model.named_parameters():
            if param.requires_grad and param.grad is not None:
                self.grad_backup[name] = param.grad.clone()

    def restore_grad(self):
        for name, param in self.model.named_parameters():
            if param.requires_grad and param.grad is not None:
                param.grad = self.grad_backup[name]

# 初始化
fgm = FGM(model)
for batch_input, batch_label in data:
  # 正常训练
  loss = model(batch_input, batch_label)
  loss.backward() # 反向传播，得到正常的grad
  # 对抗训练
  fgm.attack() # embedding被修改了
  # optimizer.zero_grad() # 如果不想累加梯度，就把这里的注释取消
  loss_sum = model(batch_input, batch_label)
  loss_sum.backward() # 反向传播，在正常的grad基础上，累加对抗训练的梯度
  fgm.restore() # 恢复Embedding的参数
  # 梯度下降，更新参数
  optimizer.step()
  optimizer.zero_grad()

1. https://wmathor.com/index.php/archives/1537/

赏

谢谢你请我吃糖果

支付宝