From 2ab6d7012e0cf20706daf5ec156c26c5f6b73eb1 Mon Sep 17 00:00:00 2001
From: "ben.guo" <909336740@qq.com>
Date: Thu, 25 Apr 2024 18:09:34 +0800
Subject: [PATCH] =?UTF-8?q?Add.=20=E5=A2=9E=E5=8A=A0=E7=AC=AC=E4=B8=80?=
 =?UTF-8?q?=E7=AB=A0=E7=BB=93=E6=9D=9F=E8=AF=AD?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 .DS_Store                                     | Bin 10244 -> 10244 bytes
 .../第一章——Transformer网络架构.md |   4 ++++
 2 files changed, 4 insertions(+)

diff --git a/.DS_Store b/.DS_Store
index 1fbb0fd6d87b3e272a2580ceb629eb602a30aac8..7b781ddd5004bce3724fe480dfc01c151db7fc39 100644
GIT binary patch
delta 20
ccmZn(XbISmDlmDMaK_|!q7IvR1;2{{09s!NN&o-=

delta 24
gcmZn(XbISmD!|A-dAD#hBgf>&qTZW%1;2{{0BK(cPXGV_

diff --git a/人人都能看懂的Transformer/第一章——Transformer网络架构.md b/人人都能看懂的Transformer/第一章——Transformer网络架构.md
index 52b60df..f30c44f 100644
--- a/人人都能看懂的Transformer/第一章——Transformer网络架构.md
+++ b/人人都能看懂的Transformer/第一章——Transformer网络架构.md
@@ -182,3 +182,7 @@ Add & Norm的过程可以理解为相同位置元素相加，再做层归一化
 > WHY：归一化到0-1区间，便于比较和处理。将注意力分数转换为概率分布。
 
 可以简单理解为，前面输出的概率，会被转化成0-1的区间进行输出。
+
+
+
+至此，你已经对整个Transformer有的整体了解，我们已经是熟悉transformer的人了🎉🎉🎉