Skip to content

Commit 8f8fde8

Browse files
committed
update
1 parent 4fdc23d commit 8f8fde8

File tree

2 files changed

+11
-3
lines changed

2 files changed

+11
-3
lines changed

_posts/2025-10-27-d2l_study_notes.md

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -24,4 +24,6 @@ author: Pianfan
2424

2525
- [6. 卷积神经网络](https://pianfan.github.io/d2l_convolutional-neural-networks/)
2626

27+
- [7. 现代卷积神经网络](https://pianfan.github.io/d2l_convolutional-modern/)
28+
2729
后续章节持续更新中……

_posts/2025-11-17-d2l_convolutional-modern.md

Lines changed: 9 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -261,7 +261,13 @@ net = nn.Sequential(
261261

262262
操作:对小批量数据进行均值和方差标准化,再应用可学习的拉伸($\gamma$)和偏移($\beta$)参数
263263

264-
公式:$\mathrm{BN}(\mathbf{x}) = \boldsymbol{\gamma} \odot \frac{\mathbf{x} - \hat{\boldsymbol{\mu}}_\mathcal{B}}{\hat{\boldsymbol{\sigma}}_\mathcal{B}} + \boldsymbol{\beta}$,其中 $\hat{\boldsymbol{\mu}}_\mathcal{B}$ 为小批量均值,$\hat{\boldsymbol{\sigma}}_\mathcal{B}$ 为小批量标准差(加 $\epsilon$ 防除零)
264+
公式:
265+
266+
$$
267+
\mathrm{BN}(\mathbf{x}) = \boldsymbol{\gamma} \odot \frac{\mathbf{x} - \hat{\boldsymbol{\mu}}_\mathcal{B}}{\hat{\boldsymbol{\sigma}}_\mathcal{B}} + \boldsymbol{\beta}
268+
$$
269+
270+
其中 $\hat{\boldsymbol{\mu}}_\mathcal{B}$ 为小批量均值,$\hat{\boldsymbol{\sigma}}_\mathcal{B}$ 为小批量标准差(加 $\epsilon$ 防除零)
265271

266272
### 7.5.2. 关键特性
267273

@@ -277,9 +283,9 @@ net = nn.Sequential(
277283

278284
### 7.5.3. 实现细节
279285

280-
全连接层:在特征维度计算均值和方差,形状为 $(1, num\_features)$
286+
全连接层:在特征维度计算均值和方差,形状为 (1, num_features)
281287

282-
卷积层:在通道维度计算均值和方差(含所有空间位置),形状为 $(1, num\_features, 1, 1)$
288+
卷积层:在通道维度计算均值和方差(含所有空间位置),形状为 (1, num_features, 1, 1)
283289

284290
PyTorch 实现:
285291

0 commit comments

Comments
 (0)