-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathatom.xml
618 lines (273 loc) · 19.2 KB
/
atom.xml
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
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom">
<title>Blogs</title>
<subtitle>Welcome to bg51717's Wiki and Blog</subtitle>
<link href="https://bg51717.github.io/atom.xml" rel="self"/>
<link href="https://bg51717.github.io/"/>
<updated>2024-11-21T00:59:30.295Z</updated>
<id>https://bg51717.github.io/</id>
<author>
<name>bg51717</name>
</author>
<generator uri="https://hexo.io/">Hexo</generator>
<entry>
<title>数学计算库:SymPy</title>
<link href="https://bg51717.github.io/16846/"/>
<id>https://bg51717.github.io/16846/</id>
<published>2024-11-19T11:19:17.000Z</published>
<updated>2024-11-21T00:59:30.295Z</updated>
<summary type="html"><h2</summary>
<category term="数学" scheme="https://bg51717.github.io/categories/%E6%95%B0%E5%AD%A6/"/>
<category term="代码" scheme="https://bg51717.github.io/categories/%E6%95%B0%E5%AD%A6/%E4%BB%A3%E7%A0%81/"/>
<category term="数学" scheme="https://bg51717.github.io/tags/%E6%95%B0%E5%AD%A6/"/>
<category term="SymPy" scheme="https://bg51717.github.io/tags/SymPy/"/>
</entry>
<entry>
<title>论文阅读习惯</title>
<link href="https://bg51717.github.io/22407/"/>
<id>https://bg51717.github.io/22407/</id>
<published>2024-11-11T13:20:18.000Z</published>
<updated>2024-11-21T00:58:37.745Z</updated>
<summary type="html"><h2</summary>
<category term="科研" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/"/>
<category term="论文阅读" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB/"/>
</entry>
<entry>
<title>OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Models</title>
<link href="https://bg51717.github.io/58958/"/>
<id>https://bg51717.github.io/58958/</id>
<published>2024-11-09T09:25:51.000Z</published>
<updated>2024-11-11T13:16:01.202Z</updated>
<summary type="html"><h2</summary>
<category term="科研" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/"/>
<category term="论文阅读" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB/"/>
</entry>
<entry>
<title>OneBit: Towards Extremely Low-bit Large Language Models</title>
<link href="https://bg51717.github.io/341/"/>
<id>https://bg51717.github.io/341/</id>
<published>2024-11-08T12:33:02.000Z</published>
<updated>2024-11-11T03:20:58.488Z</updated>
<summary type="html"><h2 id="介绍">介绍</h2>
<p>OneBit属于量化方法中的量化感知训练QAT。基于BitNet的OneBit将LLM的权重矩阵量化为1位，用一种新颖的
1</summary>
<category term="科研" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/"/>
<category term="论文阅读" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="量化" scheme="https://bg51717.github.io/tags/%E9%87%8F%E5%8C%96/"/>
</entry>
<entry>
<title>矩阵奇异值分解SVD</title>
<link href="https://bg51717.github.io/33996/"/>
<id>https://bg51717.github.io/33996/</id>
<published>2024-11-08T05:26:13.000Z</published>
<updated>2024-11-08T17:00:43.626Z</updated>
<summary type="html"><h2 id="定理陈述">定理陈述</h2>
<p>对于任意 <span class="math inline">\(m \times n\)</span> 的矩阵 <span class="math</summary>
<category term="数学" scheme="https://bg51717.github.io/categories/%E6%95%B0%E5%AD%A6/"/>
<category term="线性代数" scheme="https://bg51717.github.io/categories/%E6%95%B0%E5%AD%A6/%E7%BA%BF%E6%80%A7%E4%BB%A3%E6%95%B0/"/>
<category term="矩阵" scheme="https://bg51717.github.io/tags/%E7%9F%A9%E9%98%B5/"/>
<category term="SVD" scheme="https://bg51717.github.io/tags/SVD/"/>
</entry>
<entry>
<title>梯度估计STE</title>
<link href="https://bg51717.github.io/13277/"/>
<id>https://bg51717.github.io/13277/</id>
<published>2024-11-04T12:15:47.000Z</published>
<updated>2024-11-08T12:36:20.417Z</updated>
<summary type="html"><h2 id="背景">背景</h2>
<p>反向传播是现在训练模型的重要方法，但是在部分场景下，会遇到不可微分的函数，从而导致梯度传播失败。比如量化里的取整函数。因此，需要对梯度进行估计然后反向传播。</p>
<p><strong>STE(Straight-Through</summary>
<category term="深度学习" scheme="https://bg51717.github.io/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="经典模块" scheme="https://bg51717.github.io/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E7%BB%8F%E5%85%B8%E6%A8%A1%E5%9D%97/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="梯度估计" scheme="https://bg51717.github.io/tags/%E6%A2%AF%E5%BA%A6%E4%BC%B0%E8%AE%A1/"/>
</entry>
<entry>
<title>Welcome to bg51717's Wiki and Blog</title>
<link href="https://bg51717.github.io/14261/"/>
<id>https://bg51717.github.io/14261/</id>
<published>2024-11-02T02:55:49.874Z</published>
<updated>2024-11-02T02:55:49.874Z</updated>
<summary type="html"><p>这是bg51717的个人Wiki和Blog站点，主要是把知识系统的罗列出来以及存放一些特殊bug的处理，当然也会更一些游戏或者二次元相关东西，也希望在成长的过程中可以认识许多志同道合的人。</p>
<p>本人一直认为互联网的开源是社会发展的重要原因之一，因为开源使得技术知识</summary>
</entry>
<entry>
<title>AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration</title>
<link href="https://bg51717.github.io/14592/"/>
<id>https://bg51717.github.io/14592/</id>
<published>2024-10-27T05:35:56.000Z</published>
<updated>2024-11-11T14:55:02.024Z</updated>
<summary type="html"><h2 id="介绍">介绍</h2>
<p>本篇博客介绍论文<a href="https://arxiv.org/abs/2306.00978">[2306.00978] AWQ:
Activation-aware Weight Quantization for LLM</summary>
<category term="科研" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/"/>
<category term="论文阅读" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="量化" scheme="https://bg51717.github.io/tags/%E9%87%8F%E5%8C%96/"/>
</entry>
<entry>
<title>拉格朗日乘数法解条件极值</title>
<link href="https://bg51717.github.io/49444/"/>
<id>https://bg51717.github.io/49444/</id>
<published>2024-10-24T09:10:04.000Z</published>
<updated>2024-11-02T02:55:49.880Z</updated>
<summary type="html"><h2 id="介绍">介绍</h2>
<p>求解最优化是一类十分常见且难以求解的问题，因此，考虑开一个博客系统性的介绍一下重要解法：拉格朗日乘数法（Lagrange
Multiplier</summary>
<category term="数学" scheme="https://bg51717.github.io/categories/%E6%95%B0%E5%AD%A6/"/>
<category term="微积分" scheme="https://bg51717.github.io/categories/%E6%95%B0%E5%AD%A6/%E5%BE%AE%E7%A7%AF%E5%88%86/"/>
<category term="数学" scheme="https://bg51717.github.io/tags/%E6%95%B0%E5%AD%A6/"/>
</entry>
<entry>
<title>大模型量化~GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers</title>
<link href="https://bg51717.github.io/31769/"/>
<id>https://bg51717.github.io/31769/</id>
<published>2024-10-24T08:12:19.000Z</published>
<updated>2024-11-02T02:57:00.290Z</updated>
<summary type="html"><h2</summary>
<category term="科研" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/"/>
<category term="论文阅读" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="量化" scheme="https://bg51717.github.io/tags/%E9%87%8F%E5%8C%96/"/>
</entry>
<entry>
<title>llm.int8</title>
<link href="https://bg51717.github.io/21264/"/>
<id>https://bg51717.github.io/21264/</id>
<published>2024-10-24T07:39:30.000Z</published>
<updated>2024-11-02T02:57:00.281Z</updated>
<summary type="html"><blockquote>
<p>todo：本文还在施工中.......</p>
</blockquote>
<h2</summary>
<category term="科研" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/"/>
<category term="论文阅读" scheme="https://bg51717.github.io/categories/%E7%A7%91%E7%A0%94/%E8%AE%BA%E6%96%87%E9%98%85%E8%AF%BB/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="量化" scheme="https://bg51717.github.io/tags/%E9%87%8F%E5%8C%96/"/>
</entry>
<entry>
<title>pytorch分布式-ddp</title>
<link href="https://bg51717.github.io/45014/"/>
<id>https://bg51717.github.io/45014/</id>
<published>2024-10-09T05:30:13.000Z</published>
<updated>2024-11-02T02:55:49.879Z</updated>
<summary type="html"><h1</summary>
<category term="模板" scheme="https://bg51717.github.io/categories/%E6%A8%A1%E6%9D%BF/"/>
<category term="模板" scheme="https://bg51717.github.io/tags/%E6%A8%A1%E6%9D%BF/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="PyTorch" scheme="https://bg51717.github.io/tags/PyTorch/"/>
<category term="分布式" scheme="https://bg51717.github.io/tags/%E5%88%86%E5%B8%83%E5%BC%8F/"/>
<category term="ddq" scheme="https://bg51717.github.io/tags/ddq/"/>
<category term="数据并行" scheme="https://bg51717.github.io/tags/%E6%95%B0%E6%8D%AE%E5%B9%B6%E8%A1%8C/"/>
<category term="Data Parallel" scheme="https://bg51717.github.io/tags/Data-Parallel/"/>
</entry>
<entry>
<title>hexo博客2:双主题</title>
<link href="https://bg51717.github.io/31680/"/>
<id>https://bg51717.github.io/31680/</id>
<published>2024-10-03T10:55:45.000Z</published>
<updated>2024-11-02T05:32:24.133Z</updated>
<summary type="html"><h1 id="介绍">介绍</h1>
<p>后来发现单纯的wiki风格博客可能确实有些单调了（<del>绝对不是因为我想弄二次元风格的</del>），因此在考虑以后，决定搭建一个双主题的博客，外层是个华丽一点的主题<a</summary>
<category term="SmallProjects" scheme="https://bg51717.github.io/categories/SmallProjects/"/>
<category term="博客搭建" scheme="https://bg51717.github.io/categories/SmallProjects/%E5%8D%9A%E5%AE%A2%E6%90%AD%E5%BB%BA/"/>
<category term="博客" scheme="https://bg51717.github.io/tags/%E5%8D%9A%E5%AE%A2/"/>
<category term="Hexo" scheme="https://bg51717.github.io/tags/Hexo/"/>
<category term="Wiki" scheme="https://bg51717.github.io/tags/Wiki/"/>
</entry>
<entry>
<title>使用dotbot快速同步Linux配置</title>
<link href="https://bg51717.github.io/13162/"/>
<id>https://bg51717.github.io/13162/</id>
<published>2024-10-03T07:37:15.000Z</published>
<updated>2024-11-02T02:55:49.875Z</updated>
<summary type="html"><h1</summary>
<category term="工具" scheme="https://bg51717.github.io/categories/%E5%B7%A5%E5%85%B7/"/>
<category term="dotfiles" scheme="https://bg51717.github.io/tags/dotfiles/"/>
<category term="dotbot" scheme="https://bg51717.github.io/tags/dotbot/"/>
<category term="Linux" scheme="https://bg51717.github.io/tags/Linux/"/>
</entry>
<entry>
<title>zsh+powerlevel10K优化终端使用体验</title>
<link href="https://bg51717.github.io/13107/"/>
<id>https://bg51717.github.io/13107/</id>
<published>2024-09-29T13:09:00.000Z</published>
<updated>2024-11-02T02:55:49.874Z</updated>
<summary type="html"><h1</summary>
<category term="工具" scheme="https://bg51717.github.io/categories/%E5%B7%A5%E5%85%B7/"/>
<category term="终端" scheme="https://bg51717.github.io/tags/%E7%BB%88%E7%AB%AF/"/>
<category term="oh-my-zsh" scheme="https://bg51717.github.io/tags/oh-my-zsh/"/>
<category term="powerlevel10k" scheme="https://bg51717.github.io/tags/powerlevel10k/"/>
</entry>
<entry>
<title>安卓手机配置Google</title>
<link href="https://bg51717.github.io/61294/"/>
<id>https://bg51717.github.io/61294/</id>
<published>2024-08-27T10:14:49.000Z</published>
<updated>2024-11-02T02:55:49.875Z</updated>
<summary type="html"><h2</summary>
<category term="工具" scheme="https://bg51717.github.io/categories/%E5%B7%A5%E5%85%B7/"/>
<category term="安卓" scheme="https://bg51717.github.io/tags/%E5%AE%89%E5%8D%93/"/>
<category term="Google" scheme="https://bg51717.github.io/tags/Google/"/>
</entry>
<entry>
<title>PyTorch代码转HF</title>
<link href="https://bg51717.github.io/61054/"/>
<id>https://bg51717.github.io/61054/</id>
<published>2024-08-06T10:56:13.000Z</published>
<updated>2024-12-09T16:13:26.328Z</updated>
<summary type="html"><h2</summary>
<category term="模板" scheme="https://bg51717.github.io/categories/%E6%A8%A1%E6%9D%BF/"/>
<category term="模板" scheme="https://bg51717.github.io/tags/%E6%A8%A1%E6%9D%BF/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="PyTorch" scheme="https://bg51717.github.io/tags/PyTorch/"/>
<category term="HuggingFace" scheme="https://bg51717.github.io/tags/HuggingFace/"/>
<category term="Trainer" scheme="https://bg51717.github.io/tags/Trainer/"/>
<category term="config" scheme="https://bg51717.github.io/tags/config/"/>
<category term="model" scheme="https://bg51717.github.io/tags/model/"/>
<category term="dataset" scheme="https://bg51717.github.io/tags/dataset/"/>
</entry>
<entry>
<title>随机数种子</title>
<link href="https://bg51717.github.io/7369/"/>
<id>https://bg51717.github.io/7369/</id>
<published>2024-07-09T10:02:24.000Z</published>
<updated>2024-11-02T02:55:49.881Z</updated>
<summary type="html"><h1 id="介绍">介绍</h1>
<p>在深度学习的实际项目中，为了减少随机性，增强项目的复现能力，设置固定随机数种子十分重要，因此这篇文章罗列了一些设置随机种子的方法和减少项目随机性的经验。</p>
<h1 id="通用函数">通用函数</h1>
<pre</summary>
<category term="深度学习" scheme="https://bg51717.github.io/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="工程细节" scheme="https://bg51717.github.io/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/%E5%B7%A5%E7%A8%8B%E7%BB%86%E8%8A%82/"/>
<category term="深度学习" scheme="https://bg51717.github.io/tags/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="随机数" scheme="https://bg51717.github.io/tags/%E9%9A%8F%E6%9C%BA%E6%95%B0/"/>
</entry>
<entry>
<title>vscode调试python</title>
<link href="https://bg51717.github.io/30403/"/>
<id>https://bg51717.github.io/30403/</id>
<published>2024-04-19T14:47:51.000Z</published>
<updated>2024-11-02T02:55:49.879Z</updated>
<summary type="html"><h2 id="介绍">介绍</h2>
<p>在学习项目的过程中，很多时候需要通过调试来高效率的了解代码的执行过程，因此这里介绍下怎么使用vscode对python程序进行调试。</p>
<h2</summary>
<category term="杂项" scheme="https://bg51717.github.io/categories/%E6%9D%82%E9%A1%B9/"/>
<category term="vscode" scheme="https://bg51717.github.io/tags/vscode/"/>
<category term="python" scheme="https://bg51717.github.io/tags/python/"/>
<category term="调试" scheme="https://bg51717.github.io/tags/%E8%B0%83%E8%AF%95/"/>
<category term="debug" scheme="https://bg51717.github.io/tags/debug/"/>
</entry>
<entry>
<title>nlp常用排行榜</title>
<link href="https://bg51717.github.io/63314/"/>
<id>https://bg51717.github.io/63314/</id>
<published>2024-04-05T06:56:57.000Z</published>
<updated>2024-11-02T02:55:49.890Z</updated>
<summary type="html"><h2 id="介绍">介绍</h2>
<p>在工作和学习的时候发现，很多时候挑选合适的模型和数据集等也是一个重要且麻烦的过程。发现有很多相关的评测的排行榜，根据这些实时更新的排行榜，可以辅助我们进行选择模型等前期工作。</p>
<p><a</summary>
<category term="深度学习" scheme="https://bg51717.github.io/categories/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0/"/>
<category term="自然语言处理" scheme="https://bg51717.github.io/categories/%E8%87%AA%E7%84%B6%E8%AF%AD%E8%A8%80%E5%A4%84%E7%90%86/"/>
</entry>
</feed>