-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.xml
More file actions
236 lines (236 loc) · 16.1 KB
/
index.xml
File metadata and controls
236 lines (236 loc) · 16.1 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
<title>1Feng 的技术文章</title>
<link>http://1feng.github.io/</link>
<description>Recent content on 1Feng 的技术文章</description>
<generator>Hugo -- gohugo.io</generator>
<language>en</language>
<lastBuildDate>Fri, 02 Feb 2024 11:47:02 +0800</lastBuildDate>
<atom:link href="http://1feng.github.io/index.xml" rel="self" type="application/rss+xml" />
<item>
<title>组织管理中流程规则的利弊</title>
<link>http://1feng.github.io/post/2024-02-02-process-and-rule/</link>
<pubDate>Fri, 02 Feb 2024 11:47:02 +0800</pubDate>
<guid>http://1feng.github.io/post/2024-02-02-process-and-rule/</guid>
<description>字节的观点 最近读到腾讯深网关于字节的一篇报道,里面提到到了字节高层对组织管理流程规则的看法,整体偏负向。 张一鸣早在2016年就公开表达过自己</description>
</item>
<item>
<title>SAM 和 Visual Prompting 对 Data-Centric AI 的启示</title>
<link>http://1feng.github.io/post/2023-05-01-sam-and-visual-prompting/</link>
<pubDate>Sat, 20 May 2023 20:12:08 +0800</pubDate>
<guid>http://1feng.github.io/post/2023-05-01-sam-and-visual-prompting/</guid>
<description>整个 23 年上半年,NLP 领域的变化可谓如火如荼,那 CV 方向尤其涉及模型开发的工程架构领域发生了哪些变化?因此有了下面两个简单的调研: Ⅰ、Segm</description>
</item>
<item>
<title>Prompt Engineering 是一种什么技术?</title>
<link>http://1feng.github.io/post/2023-05-01-prompt-engineering/</link>
<pubDate>Mon, 01 May 2023 18:57:45 +0800</pubDate>
<guid>http://1feng.github.io/post/2023-05-01-prompt-engineering/</guid>
<description>本文其实是关于 Pre-Train + Prompt 在 NLP 领域的实践综述,这里的信息全部来自论文 《Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing》,这是一篇</description>
</item>
<item>
<title>后 ChatGPT 时代 Data-Centric AI 的最新思想和方法框架</title>
<link>http://1feng.github.io/post/2023-05-01-data-centric-ai/</link>
<pubDate>Mon, 01 May 2023 18:50:45 +0800</pubDate>
<guid>http://1feng.github.io/post/2023-05-01-data-centric-ai/</guid>
<description>本文内容主要总结自 23 年 3 月份的一篇论文 《Data-centric Artificial Intelligence: A Survey》。作者从必要性、目标、一般方法对 Data-centric 的思想进行了框架性的综</description>
</item>
<item>
<title>《Machine Learning Operations (MLOps): Overview, Definition, and Architecture》 summary</title>
<link>http://1feng.github.io/post/2022-08-05-mlops/</link>
<pubDate>Fri, 05 Aug 2022 15:02:44 +0800</pubDate>
<guid>http://1feng.github.io/post/2022-08-05-mlops/</guid>
<description>本文主要是针对以下论文的简要解读,其中掺杂了一些的个人理解 Machine Learning Operations (MLOps): Overview, Definition, and Architecture Dominik Kreuzberger, Niklas Kühl, Sebastian Hirschl May 2022 背景 作者认为机器学习项目的终极目标是能够把</description>
</item>
<item>
<title>互联网数据滥用</title>
<link>http://1feng.github.io/post/2022-03-31-data-abuse/</link>
<pubDate>Thu, 31 Mar 2022 17:00:51 +0800</pubDate>
<guid>http://1feng.github.io/post/2022-03-31-data-abuse/</guid>
<description>数据泄露可以造成的风险大家普遍还是概念比较清晰的。目前监管除了要求数据具备保密性之外,对数据的经营性滥用也有各种要求,那滥用究竟意味着什么呢</description>
</item>
<item>
<title>Read-Only 的 linearizability</title>
<link>http://1feng.github.io/post/2017-06-16-smarter/</link>
<pubDate>Fri, 16 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-16-smarter/</guid>
<description>《Paxos Replicated State Machines as the Basis of a High-Performance Data Store》 介绍了使用了paxos算法进行副本同步,这里仅总结如何保证read-only操作的lineari</description>
</item>
<item>
<title>线性一致性</title>
<link>http://1feng.github.io/post/2017-06-15-linearizability/</link>
<pubDate>Thu, 15 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-15-linearizability/</guid>
<description>Introduce 所谓的linearizability其目的在于描述系统的数据,对外看起来就像只有一份,所有针对这部分数据的操作都是原子(Concurren</description>
</item>
<item>
<title>弱一致性</title>
<link>http://1feng.github.io/post/2017-06-14-weak-consistency/</link>
<pubDate>Wed, 14 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-14-weak-consistency/</guid>
<description>Introduce 对于CAP而言,partition-tolerant是客观的必须要做的,不然不能称之为分布式系统;而consistent和availabl</description>
</item>
<item>
<title>zab 算法总结</title>
<link>http://1feng.github.io/post/2017-06-13-zab/</link>
<pubDate>Tue, 13 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-13-zab/</guid>
<description>Summary zab是Yahoo提出的leader-base的一致性协议,由于raft晚于该协议猜测raft中有借鉴该协议的一些思想 此文仅总结理解的一些</description>
</item>
<item>
<title>Paxos made simple</title>
<link>http://1feng.github.io/post/2017-06-12-paxos-made-simple/</link>
<pubDate>Mon, 12 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-12-paxos-made-simple/</guid>
<description>Summary paxos算法的的核心思想是“与其预测未来,不如限制当下”,即通过保证当前的操作,来一步一步达到预期 Theory 要求 Safety: 只有一个被提议的value被选</description>
</item>
<item>
<title>三阶段提交</title>
<link>http://1feng.github.io/post/2017-06-11-3pc/</link>
<pubDate>Sun, 11 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-11-3pc/</guid>
<description>Why 1983年由Dale Skeen 和 Michael Stonebraker提出了3PC协议来解决2PC阻塞的问题 #What 3PC(two-phase-commit)其实就是</description>
</item>
<item>
<title>两阶段提交</title>
<link>http://1feng.github.io/post/2017-06-10-2pc/</link>
<pubDate>Sat, 10 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-10-2pc/</guid>
<description>Why 针对数据库事务ACID-Atomicity,单机可以使用write-ahead-log实现1PC(one-phase-commit)即可,</description>
</item>
<item>
<title>分布式系统的正确性</title>
<link>http://1feng.github.io/post/2017-06-09-correctness/</link>
<pubDate>Fri, 09 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-09-correctness/</guid>
<description>Introduce 一般正确性的证明标准有两个,分别是safety properties 和 liveness properites Safety Properites 通常safety properites是指:“bad things never happen”。 举例 例如互</description>
</item>
<item>
<title>CAP 问题</title>
<link>http://1feng.github.io/post/2017-06-08-cap/</link>
<pubDate>Thu, 08 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-08-cap/</guid>
<description>Introduce 于2002年提出的CAP理论(三选二的方式来评估分布式系统)确实为分布式系统领域的发展提供了指导价值,但是就今天而言,这套理论已经意义微小</description>
</item>
<item>
<title>What is ACID</title>
<link>http://1feng.github.io/post/2017-06-07-acid/</link>
<pubDate>Wed, 07 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-07-acid/</guid>
<description>What Atomicity 描述: 一个事务包含一系列的操作,这一系列的操作都成功,则意味着事务执行成功;一旦执行过程中发生故障(fault),数据库需要放弃整个事务</description>
</item>
<item>
<title>Unreliable Network</title>
<link>http://1feng.github.io/post/2017-06-06-unreliable-network/</link>
<pubDate>Tue, 06 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-06-unreliable-network/</guid>
<description>Introduce 众所周知TCP是可靠的网络传输协议,但是为什么在分布式系统中又认为网络是不可靠的呢?通常有以下两点: 发送方无法确定接收方已经收到请求 发送方</description>
</item>
<item>
<title>Timing and Order</title>
<link>http://1feng.github.io/post/2017-06-05-timing-and-order/</link>
<pubDate>Mon, 05 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-05-timing-and-order/</guid>
<description>Introduce 分布式环境面临的两个主要的问题就是网络不可靠和时钟不可靠,这里主要总结时钟问题 Physical Clocks 我们日常使用的计算机和服务器的物理时钟都是使用的石英(q</description>
</item>
<item>
<title>Vector clock summary</title>
<link>http://1feng.github.io/post/2017-06-05-vector-clock/</link>
<pubDate>Mon, 05 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-05-vector-clock/</guid>
<description>Happend before 用→来表示hanppend before,对于任意event a, b 有: 如果a和b属于同一个process,并且a comes before b, 则 a → b 如果a是某个p</description>
</item>
<item>
<title>《Time, clocks, and the ordering of events in a distributed system》summary</title>
<link>http://1feng.github.io/post/2017-06-03-time-clocks-and-the-ordering-of-events-in-a-distributed-system/</link>
<pubDate>Sat, 03 Jun 2017 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2017-06-03-time-clocks-and-the-ordering-of-events-in-a-distributed-system/</guid>
<description>Happend before 用→来表示hanppend before,对于任意event a, b 有: 如果a和b属于同一个process,并且a comes before b, 则 a → b 如果a是某个p</description>
</item>
<item>
<title>leveldb 源码笔记之 Read</title>
<link>http://1feng.github.io/post/2016-09-10-leveldb-read/</link>
<pubDate>Sat, 10 Sep 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-09-10-leveldb-read/</guid>
<description>key逻辑分类 根据我们之前文章的描述,leveldb的数据存储可能存在在内存的memtable中,或者磁盘的sstalbe中,但是key的实</description>
</item>
<item>
<title>leveldb 源码笔记之 Compact</title>
<link>http://1feng.github.io/post/2016-09-06-leveldb-compact/</link>
<pubDate>Tue, 06 Sep 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-09-06-leveldb-compact/</guid>
<description>简介 leveldb中只有minor compaction 和 major compaction两种 代码中通过调用DBImpl::MaybeScheduleCompaction</description>
</item>
<item>
<title>leveldb 源码笔记之 MVCC && Manifest</title>
<link>http://1feng.github.io/post/2016-08-24-mvcc-and-manifest/</link>
<pubDate>Wed, 24 Aug 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-08-24-mvcc-and-manifest/</guid>
<description>MVCC 问题: 针对同一条记录,如果读和写在同一时间发生时,reader可能会读取到不一致或者写了一半的数据 常见解决方案 悲观锁: 最简单的方式,即通过</description>
</item>
<item>
<title>leveldb 源码笔记之 sstable</title>
<link>http://1feng.github.io/post/2016-08-22-sstable-summary/</link>
<pubDate>Mon, 22 Aug 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-08-22-sstable-summary/</guid>
<description>sstable 组成细节 如下: sstalbe 生成流程 sstable 生成时机: minor compaction immutable-memtable 中的key/value dump到磁盘,生成sstable major compaction sstable compact(level-n ss</description>
</item>
<item>
<title>leveldb源码笔记之 Write</title>
<link>http://1feng.github.io/post/2016-08-18-leveldb-write/</link>
<pubDate>Thu, 18 Aug 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-08-18-leveldb-write/</guid>
<description>插入一条K/V记录 持有Writer的线程进入Writers队列,细节如下: MakeRoomForWrite的流程图: 记录会首先写入磁盘上的b</description>
</item>
<item>
<title>Signal with mutex locked or not(译)</title>
<link>http://1feng.github.io/post/2016-07-20-signal-with-mutex-locked-or-not/</link>
<pubDate>Wed, 20 Jul 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-07-20-signal-with-mutex-locked-or-not/</guid>
<description>原文链接 介绍 当我们使用条件变量的时候,总有这样一个问题:到底该在解锁mutex之前进行sinal/broadcast,还是在之后? 什么时候进</description>
</item>
<item>
<title>Google File System 笔记</title>
<link>http://1feng.github.io/post/2016-07-08-google-file-system-summary/</link>
<pubDate>Fri, 08 Jul 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-07-08-google-file-system-summary/</guid>
<description>创新点 把组件故障当做常态,而不是异常。即,要有完备的监控,错误检测,故障恢复机制 通常文件都是巨大的,数以GB是常态 多数文件的修改是追加新数据</description>
</item>
<item>
<title>浮点数问题探究</title>
<link>http://1feng.github.io/post/2016-06-27-ieee-floating-point-summary/</link>
<pubDate>Mon, 27 Jun 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-06-27-ieee-floating-point-summary/</guid>
<description>问题 最近在使用openresty实现一些业务,业务中设计了一套二进制编码,目前为49bit。真正实现的时候发现lua里不支持(u)int64</description>
</item>
<item>
<title>Linux cpu load</title>
<link>http://1feng.github.io/post/2016-05-05-linux-cpu-load/</link>
<pubDate>Thu, 05 May 2016 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2016-05-05-linux-cpu-load/</guid>
<description>简析 工作一直是在linux环境下,经常通过top,uptime来查看当前机器的负载(load average), 但是始终对这个概念比较模糊,无</description>
</item>
<item>
<title>如何实现一个 daemon</title>
<link>http://1feng.github.io/post/2014-11-29-daemon-ho-to/</link>
<pubDate>Sat, 29 Nov 2014 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2014-11-29-daemon-ho-to/</guid>
<description>Linux下经常需要将服务程序实现为daemon(守护进程)。从而完成后台运行,开机启动,关机正常结束的功能。 需求 后台运行 :摆脱tty控制 根</description>
</item>
<item>
<title>hello world</title>
<link>http://1feng.github.io/post/2014-04-27-hello-world/</link>
<pubDate>Sun, 27 Apr 2014 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2014-04-27-hello-world/</guid>
<description>hello world
hello world
hello world
hello world</description>
</item>
<item>
<title>函数对象与std::bind</title>
<link>http://1feng.github.io/post/2014-04-27-std-bind-in-c-plus-plus/</link>
<pubDate>Sun, 27 Apr 2014 00:00:00 +0000</pubDate>
<guid>http://1feng.github.io/post/2014-04-27-std-bind-in-c-plus-plus/</guid>
<description>函数指针 主要有以下两类: 一般函数指针(普通函数指针 and static 成员函数指针) 非static成员函数指针 一般函数指针 初始化 普通函数指针 func_type func = function_name; or func_type func =</description>
</item>
</channel>
</rss>