-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathfashion_GNN.html
executable file
·1428 lines (1083 loc) · 66.6 KB
/
fashion_GNN.html
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
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<!DOCTYPE html>
<html>
<head>
<!-- Document Settings -->
<meta charset="utf-8" />
<meta http-equiv="X-UA-Compatible" content="IE=edge" />
<!-- On Post front-matter YAML, set "use_math: true" to use LaTex -->
<!-- Base Meta -->
<!-- dynamically fixing the title for tag/author pages -->
<title>GNN-based Fashion Coordinator</title>
<meta name="HandheldFriendly" content="True" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<!-- Styles'n'Scripts -->
<link rel="stylesheet" type="text/css" href="/assets/built/screen.css" />
<link rel="stylesheet" type="text/css" href="/assets/built/screen.edited.css" />
<link rel="stylesheet" type="text/css" href="/assets/built/syntax.css" />
<!-- syntax.css -->
<link rel="stylesheet" type="text/css" href="/assets/built/syntax.css" />
<!-- highlight.js -->
<link rel="stylesheet" href="//cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/default.min.css">
<style>.hljs { background: none; }</style>
<!--[if IE]>
<style>
p, ol, ul{
width: 100%;
}
blockquote{
width: 100%;
}
</style>
<![endif]-->
<!-- This tag outputs SEO meta+structured data and other important settings -->
<meta name="description" content="" />
<link rel="shortcut icon" href="http://0.0.0.0:4000/assets/built/images/favicon.jpg" type="image/png" />
<link rel="canonical" href="http://0.0.0.0:4000/fashion_GNN" />
<meta name="referrer" content="no-referrer-when-downgrade" />
<!--title below is coming from _includes/dynamic_title-->
<meta property="og:site_name" content="Darron's Devlog" />
<meta property="og:type" content="website" />
<meta property="og:title" content="GNN-based Fashion Coordinator" />
<meta property="og:description" content="1. About 1.1. Project Goal 1.2. Model Architecture 2. Load Data and Preprocess 3. Initialize a Graph Model 3.1. Generate graphs and edge data for each item categories 4. Model Training 4.1. Load Evaluation Data 4.2. Train with HinSAGE and Link Prediction Error 1. About 1.1. Project Goal Building a" />
<meta property="og:url" content="http://0.0.0.0:4000/fashion_GNN" />
<meta property="og:image" content="http://0.0.0.0:4000/assets/built/images/blog-cover1.png" />
<meta property="article:publisher" content="https://www.facebook.com/" />
<meta property="article:author" content="https://www.facebook.com/" />
<meta property="article:published_time" content="2021-09-14T15:00:00+00:00" />
<meta property="article:modified_time" content="2021-09-14T15:00:00+00:00" />
<meta property="article:tag" content="projects" />
<meta name="twitter:card" content="summary_large_image" />
<meta name="twitter:title" content="GNN-based Fashion Coordinator" />
<meta name="twitter:description" content="1. About 1.1. Project Goal 1.2. Model Architecture 2. Load Data and Preprocess 3. Initialize a Graph Model 3.1. Generate graphs and edge data for each item categories 4. Model Training 4.1. Load Evaluation Data 4.2. Train with HinSAGE and Link Prediction Error 1. About 1.1. Project Goal Building a" />
<meta name="twitter:url" content="http://0.0.0.0:4000/" />
<meta name="twitter:image" content="http://0.0.0.0:4000/assets/built/images/blog-cover1.png" />
<meta name="twitter:label1" content="Written by" />
<meta name="twitter:data1" content="Darron's Devlog" />
<meta name="twitter:label2" content="Filed under" />
<meta name="twitter:data2" content="projects" />
<meta name="twitter:site" content="@" />
<meta name="twitter:creator" content="@" />
<meta property="og:image:width" content="1400" />
<meta property="og:image:height" content="933" />
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Website",
"publisher": {
"@type": "Organization",
"name": "Darron's Devlog",
"logo": "http://0.0.0.0:4000/"
},
"url": "http://0.0.0.0:4000/fashion_GNN",
"image": {
"@type": "ImageObject",
"url": "http://0.0.0.0:4000/assets/built/images/blog-cover1.png",
"width": 2000,
"height": 666
},
"mainEntityOfPage": {
"@type": "WebPage",
"@id": "http://0.0.0.0:4000/fashion_GNN"
},
"description": "1. About 1.1. Project Goal 1.2. Model Architecture 2. Load Data and Preprocess 3. Initialize a Graph Model 3.1. Generate graphs and edge data for each item categories 4. Model Training 4.1. Load Evaluation Data 4.2. Train with HinSAGE and Link Prediction Error 1. About 1.1. Project Goal Building a"
}
</script>
<!-- <script type="text/javascript" src="https://demo.ghost.io/public/ghost-sdk.min.js?v=724281a32e"></script>
<script type="text/javascript">
ghost.init({
clientId: "ghost-frontend",
clientSecret: "f84a07a72b17"
});
</script> -->
<meta name="generator" content="Jekyll 3.6.2" />
<link rel="alternate" type="application/rss+xml" title="GNN-based Fashion Coordinator" href="/feed.xml" />
</head>
<body class="post-template">
<div class="site-wrapper">
<!-- All the main content gets inserted here, index.hbs, post.hbs, etc -->
<!-- default -->
<!-- The tag above means: insert everything in this file
into the {body} of the default.hbs template -->
<header class="site-header outer">
<div class="inner">
<nav class="site-nav">
<div class="site-nav-left">
<a class="site-nav-logo" href="/">Darron's Devlog</a>
<ul class="nav" role="menu">
<li class="nav-home" role="menuitem"><a href="/">Home</a></li>
<li class="nav-about" role="menuitem"><a href="/about/">About</a></li>
<li class="nav-projects" role="menuitem"><a href="/tag/projects/">Projects</a></li>
<li class="nav-studies" role="menuitem"><a href="/tag/studies/">Studies</a></li>
<li class="nav-blog" role="menuitem"><a href="/tag/blog/">Blog</a></li>
<li class="nav-archive" role="menuitem">
<a href="/archive.html">All Posts</a>
</li>
</ul>
</div>
<div class="site-nav-right">
<div class="social-links">
</div>
<a class="subscribe-button" href="#subscribe">Search</a>
</div>
</nav>
</div>
</header>
<!-- Everything inside the #post tags pulls data from the post -->
<!-- #post -->
<main id="site-main" class="site-main outer" role="main">
<div class="inner">
<article class="post-full tag-projects no-image">
<header class="post-full-header">
<section class="post-full-meta">
<time class="post-full-meta-date" datetime="14 September 2021">14 September 2021</time>
<span class="date-divider">/</span>
<a href='/tag/projects/'>PROJECTS</a>
</section>
<h1 class="post-full-title">GNN-based Fashion Coordinator</h1>
</header>
<!--
-->
<section class="post-full-content">
<div class="kg-card-markdown">
<ul id="markdown-toc">
<li><a href="#1-about" id="markdown-toc-1-about">1. About</a> <ul>
<li><a href="#11-project-goal" id="markdown-toc-11-project-goal">1.1. Project Goal</a></li>
<li><a href="#12-model-architecture" id="markdown-toc-12-model-architecture">1.2. Model Architecture</a></li>
</ul>
</li>
<li><a href="#2-load-data-and-preprocess" id="markdown-toc-2-load-data-and-preprocess">2. Load Data and Preprocess</a></li>
<li><a href="#3-initialize-a-graph-model" id="markdown-toc-3-initialize-a-graph-model">3. Initialize a Graph Model</a> <ul>
<li><a href="#31-generate-graphs-and-edge-data-for-each-item-categories" id="markdown-toc-31-generate-graphs-and-edge-data-for-each-item-categories">3.1. Generate graphs and edge data for each item categories</a></li>
</ul>
</li>
<li><a href="#4-model-training" id="markdown-toc-4-model-training">4. Model Training</a> <ul>
<li><a href="#41-load-evaluation-data" id="markdown-toc-41-load-evaluation-data">4.1. Load Evaluation Data</a></li>
<li><a href="#42-train-with-hinsage-and-link-prediction-error" id="markdown-toc-42-train-with-hinsage-and-link-prediction-error">4.2. Train with HinSAGE and Link Prediction Error</a></li>
</ul>
</li>
</ul>
<h1 id="1-about">1. About</h1>
<h2 id="11-project-goal">1.1. Project Goal</h2>
<ul>
<li>
<p>Building a <strong>“Heterogeneous GNN model”</strong> using networkx and stellargraph<br />
as a submission for <em>Fashion-how Challenge, ETRI, 2021.</em></p>
</li>
<li>
<p>Running GNN model on a cuda:GPU environment <br />
Windows 11 > docker - ubuntu kernel > CUDA on WSL</p>
</li>
<li>
<p>Data Reference:<br />
<em>Euisok Chung at al., “Dataset for Interactive Recommendation System”, HCLT-2020</em></p>
</li>
</ul>
<h2 id="12-model-architecture">1.2. Model Architecture</h2>
<p><img src="/assets/images/architecture_gnn.png" alt="png" /></p>
<h1 id="2-load-data-and-preprocess">2. Load Data and Preprocess</h1>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="c1"># a sample of raw metadata DB
</span><span class="k">print</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">name</span><span class="p">),</span><span class="nb">len</span><span class="p">(</span><span class="n">data_item</span><span class="p">))</span>
<span class="k">print</span><span class="p">(</span><span class="n">name</span><span class="p">[</span><span class="mi">0</span><span class="p">],</span><span class="n">data_item</span><span class="p">[:</span><span class="mi">4</span><span class="p">])</span> <span class="c1"># 4 descriptions for each item
</span></code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code>2607 10428
BL-001
['단추 여밈 의 전체 오픈형 스탠드 칼라 와 브이넥 네크라인 의 결합 스타일 손목 까지 내려오 는 일자형 소매 여유로운 핏 어깨 에서 허리 까지 세로 절개 에 풍성 한 러플 장식 와이드 커프스',
'면 100% 구김 이 가 기 쉬운 드라이 클리닝 권장',
'시원_해 보이 는 소라색 SKY BLUE 단색 의 깔끔_한 느낌',
'여성 스러운 페미닌 한 세련 된 사랑 스러운 깔끔_한 오피스 룩 로맨틱 한 데이트 룩 포멀 한 이미지 단정 한 오피스 걸 룩 이미지']
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="p">[</span><span class="n">cat</span><span class="p">.</span><span class="n">shape</span> <span class="k">for</span> <span class="n">cat</span> <span class="ow">in</span> <span class="n">g_model</span><span class="p">.</span><span class="n">_metadata</span><span class="p">]</span> <span class="c1"># vectorized with pre-trained subword embedding
# [(1162, 512), (673, 512), (641, 512), (131, 512)]
</span></code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="p">[</span><span class="n">cat</span><span class="p">.</span><span class="n">shape</span> <span class="k">for</span> <span class="n">cat</span> <span class="ow">in</span> <span class="n">g_model</span><span class="p">.</span><span class="n">_feats</span><span class="p">]</span> <span class="c1"># vecotrized image features
# [(1162, 4096), (673, 4096), (641, 4096), (131, 4096)]
</span></code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="c1"># in 4 categories, each items has 2 vectorized data, 512-length metadata & 4096-length imagedata
</span><span class="k">print</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">slot_name</span><span class="p">))</span> <span class="c1"># 4
</span><span class="k">print</span><span class="p">(</span><span class="n">slot_name</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">0</span><span class="p">],</span> <span class="n">slot_item</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">0</span><span class="p">].</span><span class="n">shape</span><span class="p">,</span> <span class="n">slot_feat</span><span class="p">[</span><span class="mi">0</span><span class="p">][</span><span class="mi">0</span><span class="p">].</span><span class="n">shape</span><span class="p">)</span>
<span class="c1"># CD-001 (512,) (4096,)
</span></code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="p">[</span><span class="n">cat</span><span class="p">.</span><span class="n">shape</span> <span class="k">for</span> <span class="n">cat</span> <span class="ow">in</span> <span class="n">g_model</span><span class="p">.</span><span class="n">_meta_similarities</span><span class="p">]</span>
<span class="c1"># [(1162, 1162), (673, 673), (641, 641), (131, 131)]
</span></code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="kn">from</span> <span class="nn">file_io_edit</span> <span class="kn">import</span> <span class="n">_load_trn_dialog</span>
<span class="n">dialog</span><span class="p">,</span> <span class="n">coordi</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">delim_dlg</span><span class="p">,</span> <span class="n">delim_crd</span><span class="p">,</span> <span class="n">delim_rwd</span> <span class="o">=</span> \
<span class="n">_load_trn_dialog</span><span class="p">(</span><span class="n">in_file_dialog</span><span class="p">)</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="k">print</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">dialog</span><span class="p">))</span> <span class="c1"># all sentences in full dialogs
</span><span class="k">print</span><span class="p">(</span><span class="n">dialog</span><span class="p">[:</span><span class="mi">21</span><span class="p">])</span> <span class="c1"># example of the first dialog
</span></code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code>92444
['어서 오 세 요 코디 봇 입 니다 무엇 을 도와 드릴_까 요',
'처음 대학교 들어가 는데 입 을 옷 코디 해 주 세 요',
'신입생 코디 에 어울리 게 화사 한 스웨터 를 추천_해 드릴_게 요',
'이 옷 에 어울리 는 치마 로 추천_해 주 세 요',
'고객 님 의 키 사이즈 에 맞추 면 이런 옷 도 잘 어울리 실 것 같_은데 어떠 신가 요',
'제 가 키 가 작_아서 짧 은 치마 로 추천_해 주 세 요',
'상의 색상 과 도 매칭 이 잘 어울리 는 짧 은 치마 입 니다',
'어두운 계열 은 없 나 요',
'언밸런스 한 컷팅 으로 세련미 를 돋보이 게_하 는 치마 인데 마음 에 드 시_나 요',
'나쁘 지_않 네 요 외투 도 추천_해 주 시 겠_어 요',
'요즘 계절 에는 가디건 이나 자켓 을 걸치기 에 좋_은데 특정 종류 로 원하 는 게 있 으신가 요',
'트렌치 코트 종류 로 추천_해 주 세 요',
'이너 색상 과 무난_하 게 잘 어울릴 트렌치 코트 입 니다',
'신발 도 추천_해 주 세 요',
'운동화 나 구두 중 어떤 걸 선호_하 시_나 요',
'운동화 로 추천_해 주 세 요',
'어떤 스타일 과 도 무난_하 게 잘 어울리 는 기본 아이템 입 니다',
'맘 에 드_네 요 전체 코디샷 볼_수 있 나 요',
'네 지금 까지 제안 해 드린 아이템 으로 전체 코디샷 을 제안 해 드립 니다 마음 에 드 시_나 요',
'네 마음 에 드_네 요 감사_합 니다',
'마음 에 드 신_다 니 다행 입 니다 이용_해 주 셔 서 감사_합 니다']
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="kn">from</span> <span class="nn">file_io_edit</span> <span class="kn">import</span> <span class="n">_episode_slice</span>
<span class="n">dialog</span> <span class="o">=</span> <span class="n">_episode_slice</span><span class="p">(</span><span class="n">dialog</span><span class="p">,</span> <span class="n">delim_dlg</span><span class="p">)</span>
<span class="n">coordi</span> <span class="o">=</span> <span class="n">_episode_slice</span><span class="p">(</span><span class="n">coordi</span><span class="p">,</span> <span class="n">delim_crd</span><span class="p">)</span>
<span class="n">reward</span> <span class="o">=</span> <span class="n">_episode_slice</span><span class="p">(</span><span class="n">reward</span><span class="p">,</span> <span class="n">delim_rwd</span><span class="p">)</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="k">print</span><span class="p">(</span><span class="nb">len</span><span class="p">(</span><span class="n">dialog</span><span class="p">))</span>
<span class="k">print</span><span class="p">(</span><span class="n">dialog</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span> <span class="c1"># cut for each dialogs
</span></code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code>7236
['어서 오 세 요 코디 봇 입 니다 무엇 을 도와 드릴_까 요',
'처음 대학교 들어가 는데 입 을 옷 코디 해 주 세 요',
'신입생 코디 에 어울리 게 화사 한 스웨터 를 추천_해 드릴_게 요',
'이 옷 에 어울리 는 치마 로 추천_해 주 세 요',
'고객 님 의 키 사이즈 에 맞추 면 이런 옷 도 잘 어울리 실 것 같_은데 어떠 신가 요',
'제 가 키 가 작_아서 짧 은 치마 로 추천_해 주 세 요',
'상의 색상 과 도 매칭 이 잘 어울리 는 짧 은 치마 입 니다',
'어두운 계열 은 없 나 요',
'언밸런스 한 컷팅 으로 세련미 를 돋보이 게_하 는 치마 인데 마음 에 드 시_나 요',
'나쁘 지_않 네 요 외투 도 추천_해 주 시 겠_어 요',
'요즘 계절 에는 가디건 이나 자켓 을 걸치기 에 좋_은데 특정 종류 로 원하 는 게 있 으신가 요',
'트렌치 코트 종류 로 추천_해 주 세 요',
'이너 색상 과 무난_하 게 잘 어울릴 트렌치 코트 입 니다',
'신발 도 추천_해 주 세 요',
'운동화 나 구두 중 어떤 걸 선호_하 시_나 요',
'운동화 로 추천_해 주 세 요',
'어떤 스타일 과 도 무난_하 게 잘 어울리 는 기본 아이템 입 니다',
'맘 에 드_네 요 전체 코디샷 볼_수 있 나 요',
'네 지금 까지 제안 해 드린 아이템 으로 전체 코디샷 을 제안 해 드립 니다 마음 에 드 시_나 요',
'네 마음 에 드_네 요 감사_합 니다',
'마음 에 드 신_다 니 다행 입 니다 이용_해 주 셔 서 감사_합 니다']
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="c1"># vectorized dialog data
</span><span class="n">g_model</span><span class="p">.</span><span class="n">_mem_trn_dlg</span><span class="p">.</span><span class="n">shape</span> <span class="c1"># (7236, 2048)
</span></code></pre></div></div>
<h1 id="3-initialize-a-graph-model">3. Initialize a Graph Model</h1>
<ul>
<li>Procedure of loading and preprocessing data included</li>
</ul>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="kn">import</span> <span class="nn">random</span>
<span class="n">random</span><span class="p">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">2021</span><span class="p">)</span>
<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="n">np</span>
<span class="n">np</span><span class="p">.</span><span class="n">random</span><span class="p">.</span><span class="n">seed</span><span class="p">(</span><span class="mi">2021</span><span class="p">)</span>
<span class="kn">import</span> <span class="nn">tensorflow</span> <span class="k">as</span> <span class="n">tf</span>
<span class="n">tf</span><span class="p">.</span><span class="n">random</span><span class="p">.</span><span class="n">set_seed</span><span class="p">(</span><span class="mi">2021</span><span class="p">)</span>
<span class="n">tf</span><span class="p">.</span><span class="n">device</span><span class="p">(</span><span class="s">'/device:GPU:0'</span><span class="p">)</span>
<span class="kn">import</span> <span class="nn">networkx</span> <span class="k">as</span> <span class="n">nx</span>
<span class="kn">import</span> <span class="nn">stellargraph</span> <span class="k">as</span> <span class="n">sg</span>
<span class="kn">import</span> <span class="nn">argparse</span>
<span class="kn">from</span> <span class="nn">graph_model</span> <span class="kn">import</span> <span class="o">*</span>
<span class="kn">import</span> <span class="nn">os</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span> <span class="o">=</span> <span class="n">graph_model</span><span class="p">(</span><span class="n">args</span><span class="p">)</span>
</code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code><Initialize subword embedding>
loading= ./sstm_v0p5_deploy/sstm_v4p49_np_final_n36134_d128_r_eng_upper.dat
<Make metadata>
loading fashion item metadata
vectorizing data
<Make input & output data>
loading dialog DB
# of dialog: 7236 sets
vectorizing data
memorizing data
<Make input & output data>
loading dialog DB
# of dialog: 200 sets
vectorizing data
memorizing data
</code></pre></div></div>
<h2 id="31-generate-graphs-and-edge-data-for-each-item-categories">3.1. Generate graphs and edge data for each item categories</h2>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span><span class="p">.</span><span class="n">_graph_cats</span>
</code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code>(<stellargraph.core.graph.StellarGraph at 0x7f58a4f11bb0>,
<stellargraph.core.graph.StellarGraph at 0x7f58a4f11f10>,
<stellargraph.core.graph.StellarGraph at 0x7f58a4f11c40>,
<stellargraph.core.graph.StellarGraph at 0x7f58a4758af0>)
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="k">for</span> <span class="n">g</span> <span class="ow">in</span> <span class="n">g_model</span><span class="p">.</span><span class="n">_graph_cats</span><span class="p">:</span>
<span class="k">print</span><span class="p">(</span><span class="n">g</span><span class="p">.</span><span class="n">info</span><span class="p">())</span>
</code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code>StellarGraph: Undirected multigraph
Nodes: 8398, Edges: 9222
Node types:
dialog: [7236]
Features: float32 vector, length 2048
Edge types: dialog-default->item
item: [1162]
Features: float32 vector, length 4608
Edge types: item-default->dialog
Edge types:
dialog-default->item: [9222]
Weights: range=[0.5, 1], mean=0.607623, std=0.205508
Features: none
StellarGraph: Undirected multigraph
Nodes: 7909, Edges: 9022
Node types:
dialog: [7236]
Features: float32 vector, length 2048
Edge types: dialog-default->item
item: [673]
Features: float32 vector, length 4608
Edge types: item-default->dialog
Edge types:
dialog-default->item: [9022]
Weights: range=[0.5, 1], mean=0.607515, std=0.205433
Features: none
StellarGraph: Undirected multigraph
Nodes: 7877, Edges: 11549
Node types:
dialog: [7236]
Features: float32 vector, length 2048
Edge types: dialog-default->item
item: [641]
Features: float32 vector, length 4608
Edge types: item-default->dialog
Edge types:
dialog-default->item: [11549]
Weights: range=[0.5, 1], mean=0.613733, std=0.209607
Features: none
StellarGraph: Undirected multigraph
Nodes: 7367, Edges: 9757
Node types:
dialog: [7236]
Features: float32 vector, length 2048
Edge types: dialog-default->item
item: [131]
Features: float32 vector, length 4608
Edge types: item-default->dialog
Edge types:
dialog-default->item: [9757]
Weights: range=[0.5, 1], mean=0.618479, std=0.212619
Features: none
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span><span class="p">.</span><span class="n">_graph_datas</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge">
<table>
<thead>
<tr style="text-align: right;">
<th></th>
<th>dialog</th>
<th>item</th>
<th>label</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>d_2</td>
<td>o_0</td>
<td>1.0</td>
</tr>
<tr>
<th>1</th>
<td>d_53</td>
<td>o_0</td>
<td>1.0</td>
</tr>
<tr>
<th>2</th>
<td>d_133</td>
<td>o_0</td>
<td>0.5</td>
</tr>
<tr>
<th>3</th>
<td>d_337</td>
<td>o_0</td>
<td>0.5</td>
</tr>
<tr>
<th>4</th>
<td>d_437</td>
<td>o_0</td>
<td>0.5</td>
</tr>
<tr>
<th>...</th>
<td>...</td>
<td>...</td>
<td>...</td>
</tr>
<tr>
<th>9217</th>
<td>d_5649</td>
<td>o_1160</td>
<td>0.5</td>
</tr>
<tr>
<th>9218</th>
<td>d_5938</td>
<td>o_1160</td>
<td>0.5</td>
</tr>
<tr>
<th>9219</th>
<td>d_5972</td>
<td>o_1160</td>
<td>1.0</td>
</tr>
<tr>
<th>9220</th>
<td>d_6439</td>
<td>o_1160</td>
<td>0.5</td>
</tr>
<tr>
<th>9221</th>
<td>d_6788</td>
<td>o_1160</td>
<td>0.5</td>
</tr>
</tbody>
</table>
<p>9222 rows × 3 columns</p>
</div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span><span class="p">.</span><span class="n">_graph_datas</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge">
<table>
<thead>
<tr style="text-align: right;">
<th></th>
<th>dialog</th>
<th>item</th>
<th>label</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>d_514</td>
<td>t_0</td>
<td>0.5</td>
</tr>
<tr>
<th>1</th>
<td>d_560</td>
<td>t_0</td>
<td>0.5</td>
</tr>
<tr>
<th>2</th>
<td>d_815</td>
<td>t_0</td>
<td>0.5</td>
</tr>
<tr>
<th>3</th>
<td>d_839</td>
<td>t_0</td>
<td>0.5</td>
</tr>
<tr>
<th>4</th>
<td>d_884</td>
<td>t_0</td>
<td>0.5</td>
</tr>
<tr>
<th>...</th>
<td>...</td>
<td>...</td>
<td>...</td>
</tr>
<tr>
<th>9017</th>
<td>d_5905</td>
<td>t_672</td>
<td>1.0</td>
</tr>
<tr>
<th>9018</th>
<td>d_5919</td>
<td>t_672</td>
<td>0.5</td>
</tr>
<tr>
<th>9019</th>
<td>d_5999</td>
<td>t_672</td>
<td>0.5</td>
</tr>
<tr>
<th>9020</th>
<td>d_6745</td>
<td>t_672</td>
<td>1.0</td>
</tr>
<tr>
<th>9021</th>
<td>d_6746</td>
<td>t_672</td>
<td>0.5</td>
</tr>
</tbody>
</table>
<p>9022 rows × 3 columns</p>
</div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span><span class="p">.</span><span class="n">_graph_datas</span><span class="p">[</span><span class="mi">2</span><span class="p">]</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge">
<table>
<thead>
<tr style="text-align: right;">
<th></th>
<th>dialog</th>
<th>item</th>
<th>label</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>d_34</td>
<td>b_0</td>
<td>0.5</td>
</tr>
<tr>
<th>1</th>
<td>d_225</td>
<td>b_0</td>
<td>0.5</td>
</tr>
<tr>
<th>2</th>
<td>d_272</td>
<td>b_0</td>
<td>1.0</td>
</tr>
<tr>
<th>3</th>
<td>d_302</td>
<td>b_0</td>
<td>1.0</td>
</tr>
<tr>
<th>4</th>
<td>d_380</td>
<td>b_0</td>
<td>0.5</td>
</tr>
<tr>
<th>...</th>
<td>...</td>
<td>...</td>
<td>...</td>
</tr>
<tr>
<th>11544</th>
<td>d_6239</td>
<td>b_637</td>
<td>0.5</td>
</tr>
<tr>
<th>11545</th>
<td>d_6379</td>
<td>b_637</td>
<td>0.5</td>
</tr>
<tr>
<th>11546</th>
<td>d_5506</td>
<td>b_638</td>
<td>0.5</td>
</tr>
<tr>
<th>11547</th>
<td>d_5576</td>
<td>b_638</td>
<td>0.5</td>
</tr>
<tr>
<th>11548</th>
<td>d_5097</td>
<td>b_639</td>
<td>1.0</td>
</tr>
</tbody>
</table>
<p>11549 rows × 3 columns</p>
</div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span><span class="p">.</span><span class="n">_graph_datas</span><span class="p">[</span><span class="mi">3</span><span class="p">]</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge">
<table>
<thead>
<tr style="text-align: right;">
<th></th>
<th>dialog</th>
<th>item</th>
<th>label</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>d_47</td>
<td>s_0</td>
<td>0.5</td>
</tr>
<tr>
<th>1</th>
<td>d_52</td>
<td>s_0</td>
<td>0.5</td>
</tr>
<tr>
<th>2</th>
<td>d_74</td>
<td>s_0</td>
<td>0.5</td>
</tr>
<tr>
<th>3</th>
<td>d_124</td>
<td>s_0</td>
<td>0.5</td>
</tr>
<tr>
<th>4</th>
<td>d_175</td>
<td>s_0</td>
<td>0.5</td>
</tr>
<tr>
<th>...</th>
<td>...</td>
<td>...</td>
<td>...</td>
</tr>
<tr>
<th>9752</th>
<td>d_6643</td>
<td>s_129</td>
<td>0.5</td>
</tr>
<tr>
<th>9753</th>
<td>d_6684</td>
<td>s_129</td>
<td>0.5</td>
</tr>
<tr>
<th>9754</th>
<td>d_7050</td>
<td>s_129</td>
<td>1.0</td>
</tr>
<tr>
<th>9755</th>
<td>d_7082</td>
<td>s_129</td>
<td>0.5</td>
</tr>
<tr>
<th>9756</th>
<td>d_7124</td>
<td>s_129</td>
<td>1.0</td>
</tr>
</tbody>
</table>
<p>9757 rows × 3 columns</p>
</div>
<h1 id="4-model-training">4. Model Training</h1>
<h2 id="41-load-evaluation-data">4.1. Load Evaluation Data</h2>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">np</span><span class="p">.</span><span class="n">array</span><span class="p">(</span><span class="n">g_model</span><span class="p">.</span><span class="n">_mem_tst_dlg</span><span class="p">).</span><span class="n">shape</span> <span class="c1"># (200, 2048)
</span></code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">np</span><span class="p">.</span><span class="n">array</span><span class="p">(</span><span class="n">g_model</span><span class="p">.</span><span class="n">_tst_crd</span><span class="p">).</span><span class="n">shape</span> <span class="c1"># (200, 3, 4)
</span></code></pre></div></div>
<h2 id="42-train-with-hinsage-and-link-prediction-error">4.2. Train with HinSAGE and Link Prediction Error</h2>
<ul>
<li>work on building a base model, not fully optimized</li>
</ul>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="p">...</span>
<span class="c1"># training function
</span><span class="k">def</span> <span class="nf">_graph_train</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span><span class="n">data</span><span class="p">,</span><span class="n">data_graph</span><span class="p">):</span>
<span class="n">batch_size</span> <span class="o">=</span> <span class="mi">200</span>
<span class="n">epochs</span> <span class="o">=</span> <span class="mi">10</span>
<span class="n">train_size</span> <span class="o">=</span> <span class="mf">0.7</span>
<span class="n">test_size</span> <span class="o">=</span> <span class="mf">0.3</span>
<span class="n">num_samples</span> <span class="o">=</span> <span class="p">[</span><span class="mi">8</span><span class="p">,</span> <span class="mi">4</span><span class="p">]</span>
<span class="n">num_workers</span> <span class="o">=</span> <span class="mi">2</span>
<span class="n">edges_train</span><span class="p">,</span> <span class="n">edges_test</span> <span class="o">=</span> <span class="n">model_selection</span><span class="p">.</span><span class="n">train_test_split</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="n">train_size</span><span class="o">=</span><span class="n">train_size</span><span class="p">,</span> <span class="n">test_size</span><span class="o">=</span><span class="n">test_size</span><span class="p">)</span>
<span class="n">edgelist_train</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">edges_train</span><span class="p">[[</span><span class="s">"dialog"</span><span class="p">,</span> <span class="s">"item"</span><span class="p">]].</span><span class="n">itertuples</span><span class="p">(</span><span class="n">index</span><span class="o">=</span><span class="bp">False</span><span class="p">))</span>
<span class="n">edgelist_test</span> <span class="o">=</span> <span class="nb">list</span><span class="p">(</span><span class="n">edges_test</span><span class="p">[[</span><span class="s">"dialog"</span><span class="p">,</span> <span class="s">"item"</span><span class="p">]].</span><span class="n">itertuples</span><span class="p">(</span><span class="n">index</span><span class="o">=</span><span class="bp">False</span><span class="p">))</span>
<span class="n">labels_train</span> <span class="o">=</span> <span class="n">edges_train</span><span class="p">[</span><span class="s">"label"</span><span class="p">]</span>
<span class="n">labels_test</span> <span class="o">=</span> <span class="n">edges_test</span><span class="p">[</span><span class="s">"label"</span><span class="p">]</span>
<span class="n">generator</span> <span class="o">=</span> <span class="n">HinSAGELinkGenerator</span><span class="p">(</span>
<span class="n">data_graph</span><span class="p">,</span> <span class="n">batch_size</span><span class="p">,</span> <span class="n">num_samples</span><span class="p">,</span> <span class="n">head_node_types</span><span class="o">=</span><span class="p">[</span><span class="s">"dialog"</span><span class="p">,</span> <span class="s">"item"</span><span class="p">])</span>
<span class="n">train_gen</span> <span class="o">=</span> <span class="n">generator</span><span class="p">.</span><span class="n">flow</span><span class="p">(</span><span class="n">edgelist_train</span><span class="p">,</span> <span class="n">labels_train</span><span class="p">,</span> <span class="n">shuffle</span><span class="o">=</span><span class="bp">True</span><span class="p">)</span>
<span class="n">test_gen</span> <span class="o">=</span> <span class="n">generator</span><span class="p">.</span><span class="n">flow</span><span class="p">(</span><span class="n">edgelist_test</span><span class="p">,</span> <span class="n">labels_test</span><span class="p">)</span>
<span class="n">hinsage_layer_sizes</span> <span class="o">=</span> <span class="p">[</span><span class="mi">32</span><span class="p">,</span> <span class="mi">32</span><span class="p">]</span>
<span class="k">assert</span> <span class="nb">len</span><span class="p">(</span><span class="n">hinsage_layer_sizes</span><span class="p">)</span> <span class="o">==</span> <span class="nb">len</span><span class="p">(</span><span class="n">num_samples</span><span class="p">)</span>
<span class="n">hinsage</span> <span class="o">=</span> <span class="n">HinSAGE</span><span class="p">(</span>
<span class="n">layer_sizes</span><span class="o">=</span><span class="n">hinsage_layer_sizes</span><span class="p">,</span> <span class="n">generator</span><span class="o">=</span><span class="n">generator</span><span class="p">,</span> <span class="n">bias</span><span class="o">=</span><span class="bp">True</span><span class="p">,</span> <span class="n">dropout</span><span class="o">=</span><span class="mf">0.0</span><span class="p">)</span>
<span class="n">x_inp</span><span class="p">,</span> <span class="n">x_out</span> <span class="o">=</span> <span class="n">hinsage</span><span class="p">.</span><span class="n">in_out_tensors</span><span class="p">()</span>
<span class="n">score_prediction</span> <span class="o">=</span> <span class="n">link_regression</span><span class="p">(</span><span class="n">edge_embedding_method</span><span class="o">=</span><span class="s">"concat"</span><span class="p">)(</span><span class="n">x_out</span><span class="p">)</span>
<span class="n">model</span> <span class="o">=</span> <span class="n">Model</span><span class="p">(</span><span class="n">inputs</span><span class="o">=</span><span class="n">x_inp</span><span class="p">,</span> <span class="n">outputs</span><span class="o">=</span><span class="n">score_prediction</span><span class="p">)</span>
<span class="n">model</span><span class="p">.</span><span class="nb">compile</span><span class="p">(</span>
<span class="n">optimizer</span><span class="o">=</span><span class="n">optimizers</span><span class="p">.</span><span class="n">Adam</span><span class="p">(</span><span class="n">learning_rate</span><span class="o">=</span><span class="mf">1e-2</span><span class="p">),</span>
<span class="n">loss</span><span class="o">=</span><span class="n">losses</span><span class="p">.</span><span class="n">mean_squared_error</span><span class="p">,</span>
<span class="n">metrics</span><span class="o">=</span><span class="p">[</span><span class="n">root_mean_square_error</span><span class="p">,</span> <span class="n">metrics</span><span class="p">.</span><span class="n">mae</span><span class="p">],)</span>
<span class="n">test_metrics</span> <span class="o">=</span> <span class="n">model</span><span class="p">.</span><span class="n">evaluate</span><span class="p">(</span>
<span class="n">test_gen</span><span class="p">,</span> <span class="n">verbose</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">use_multiprocessing</span><span class="o">=</span><span class="bp">False</span><span class="p">,</span> <span class="n">workers</span><span class="o">=</span><span class="n">num_workers</span>
<span class="p">)</span>
<span class="n">history</span> <span class="o">=</span> <span class="n">model</span><span class="p">.</span><span class="n">fit</span><span class="p">(</span>
<span class="n">train_gen</span><span class="p">,</span>
<span class="n">validation_data</span><span class="o">=</span><span class="n">test_gen</span><span class="p">,</span>
<span class="n">epochs</span><span class="o">=</span><span class="n">epochs</span><span class="p">,</span>
<span class="n">verbose</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span>
<span class="n">shuffle</span><span class="o">=</span><span class="bp">False</span><span class="p">,</span>
<span class="n">use_multiprocessing</span><span class="o">=</span><span class="bp">False</span><span class="p">,</span>
<span class="n">workers</span><span class="o">=</span><span class="n">num_workers</span><span class="p">,)</span>
<span class="n">test_metrics</span> <span class="o">=</span> <span class="n">model</span><span class="p">.</span><span class="n">evaluate</span><span class="p">(</span>
<span class="n">test_gen</span><span class="p">,</span> <span class="n">use_multiprocessing</span><span class="o">=</span><span class="bp">False</span><span class="p">,</span> <span class="n">workers</span><span class="o">=</span><span class="n">num_workers</span><span class="p">,</span> <span class="n">verbose</span><span class="o">=</span><span class="mi">1</span><span class="p">)</span>
<span class="k">print</span><span class="p">(</span><span class="s">"Test Evaluation:"</span><span class="p">)</span>
<span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">val</span> <span class="ow">in</span> <span class="nb">zip</span><span class="p">(</span><span class="n">model</span><span class="p">.</span><span class="n">metrics_names</span><span class="p">,</span> <span class="n">test_metrics</span><span class="p">):</span>
<span class="k">print</span><span class="p">(</span><span class="s">"</span><span class="se">\t</span><span class="s">{}: {:0.4f}"</span><span class="p">.</span><span class="nb">format</span><span class="p">(</span><span class="n">name</span><span class="p">,</span> <span class="n">val</span><span class="p">))</span>
<span class="n">y_true</span> <span class="o">=</span> <span class="n">labels_test</span>
<span class="n">y_pred</span> <span class="o">=</span> <span class="n">model</span><span class="p">.</span><span class="n">predict</span><span class="p">(</span><span class="n">test_gen</span><span class="p">)</span>
<span class="n">y_pred_baseline</span> <span class="o">=</span> <span class="n">np</span><span class="p">.</span><span class="n">full_like</span><span class="p">(</span><span class="n">y_pred</span><span class="p">,</span> <span class="n">np</span><span class="p">.</span><span class="n">mean</span><span class="p">(</span><span class="n">y_true</span><span class="p">))</span>
<span class="p">...</span>
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="n">g_model</span><span class="p">.</span><span class="n">train</span><span class="p">()</span>
</code></pre></div></div>
<div class="language-plaintext highlighter-rouge"><div class="highlight"><pre class="highlight"><code>link_regression: using 'concat' method to combine node embeddings into edge embeddings
14/14 [==============================] - 3s 92ms/step - loss: 0.1245 - root_mean_square_error: 0.3521 - mean_absolute_error: 0.2666
Untrained model's Test Evaluation:
loss: 0.1245
root_mean_square_error: 0.3521
mean_absolute_error: 0.2666
Epoch 1/10
33/33 [==============================] - 8s 222ms/step - loss: 0.0947 - root_mean_square_error: 0.2792 - mean_absolute_error: 0.2281 - val_loss: 0.0451 - val_root_mean_square_error: 0.2123 - val_mean_absolute_error: 0.1991
Epoch 2/10
33/33 [==============================] - 7s 217ms/step - loss: 0.0441 - root_mean_square_error: 0.2093 - mean_absolute_error: 0.1761 - val_loss: 0.0427 - val_root_mean_square_error: 0.2062 - val_mean_absolute_error: 0.1558
Epoch 3/10
33/33 [==============================] - 8s 226ms/step - loss: 0.0426 - root_mean_square_error: 0.2062 - mean_absolute_error: 0.1684 - val_loss: 0.0423 - val_root_mean_square_error: 0.2053 - val_mean_absolute_error: 0.1613
Epoch 4/10
33/33 [==============================] - 8s 224ms/step - loss: 0.0425 - root_mean_square_error: 0.2053 - mean_absolute_error: 0.1683 - val_loss: 0.0422 - val_root_mean_square_error: 0.2050 - val_mean_absolute_error: 0.1620
Epoch 5/10
33/33 [==============================] - 8s 234ms/step - loss: 0.0423 - root_mean_square_error: 0.2053 - mean_absolute_error: 0.1679 - val_loss: 0.0425 - val_root_mean_square_error: 0.2058 - val_mean_absolute_error: 0.1805
Epoch 6/10
33/33 [==============================] - 8s 231ms/step - loss: 0.0423 - root_mean_square_error: 0.2057 - mean_absolute_error: 0.1682 - val_loss: 0.0420 - val_root_mean_square_error: 0.2047 - val_mean_absolute_error: 0.1751
Epoch 7/10
33/33 [==============================] - 8s 233ms/step - loss: 0.0421 - root_mean_square_error: 0.2046 - mean_absolute_error: 0.1701 - val_loss: 0.0417 - val_root_mean_square_error: 0.2039 - val_mean_absolute_error: 0.1685
Epoch 8/10
33/33 [==============================] - 8s 229ms/step - loss: 0.0421 - root_mean_square_error: 0.2050 - mean_absolute_error: 0.1662 - val_loss: 0.0415 - val_root_mean_square_error: 0.2033 - val_mean_absolute_error: 0.1686
Epoch 9/10
33/33 [==============================] - 8s 238ms/step - loss: 0.0415 - root_mean_square_error: 0.2036 - mean_absolute_error: 0.1679 - val_loss: 0.0411 - val_root_mean_square_error: 0.2025 - val_mean_absolute_error: 0.1695
Epoch 10/10
33/33 [==============================] - 8s 232ms/step - loss: 0.0411 - root_mean_square_error: 0.2023 - mean_absolute_error: 0.1668 - val_loss: 0.0409 - val_root_mean_square_error: 0.2017 - val_mean_absolute_error: 0.1535
14/14 [==============================] - 2s 121ms/step - loss: 0.0408 - root_mean_square_error: 0.2016 - mean_absolute_error: 0.1535
Test Evaluation:
loss: 0.0408
root_mean_square_error: 0.2016
mean_absolute_error: 0.1535
Mean Baseline Test set metrics:
root_mean_square_error = 0.2054230809528016
mean_absolute_error = 0.1687945779971711
Model Test set metrics:
root_mean_square_error = 0.20211308444911932
mean_absolute_error = 0.15357324988306925
...
...
...
Done training
</code></pre></div></div>
<div class="language-python highlighter-rouge"><div class="highlight"><pre class="highlight"><code><span class="kn">from</span> <span class="nn">numba</span> <span class="kn">import</span> <span class="n">cuda</span>
<span class="n">device</span> <span class="o">=</span> <span class="n">cuda</span><span class="p">.</span><span class="n">get_current_device</span><span class="p">()</span>
<span class="n">device</span><span class="p">.</span><span class="n">reset</span><span class="p">()</span>
</code></pre></div></div>
</div>
</section>
<!-- Email subscribe form at the bottom of the page -->
<!--
<section class="subscribe-form">
<h3 class="subscribe-form-title">Subscribe to Darron's Devlog</h3>
<p>Get the latest posts delivered right to your inbox</p>
<span id="searchform" method="post" action="/subscribe/" class="">
<input class="confirm" type="hidden" name="confirm" />
<input class="location" type="hidden" name="location" />
<input class="referrer" type="hidden" name="referrer" />
<div class="form-group">
<input class="subscribe-email" onkeyup="myFunc()"
id="searchtext" type="text" name="searchtext"
placeholder="Search..." />
</div>
<script type="text/javascript">
function myFunc() {
if(event.keyCode == 13) {
var url = encodeURIComponent($("#searchtext").val());
location.href = "/search.html?query=" + url;
}
}
</script>
</span>
</section>
-->
<footer class="post-full-footer">
<!-- Everything inside the #author tags pulls data from the author -->
<!-- #author-->
<!-- /author -->
</footer>
<!-- If you use Disqus comments, just uncomment this block.
The only thing you need to change is "test-apkdzgmqhj" - which
should be replaced with your own Disqus site-id. -->
<section class="post-full-comments">
<div id="disqus_thread"></div>
<script>
var disqus_config = function () {
var this_page_url = 'http://0.0.0.0:4000/fashion_GNN';
var this_page_identifier = '/fashion_GNN';
var this_page_title = 'GNN-based Fashion Coordinator';
};
(function() {
var d = document, s = d.createElement('script');
s.src = 'https://.disqus.com/embed.js';
s.setAttribute('data-timestamp', +new Date());
(d.head || d.body).appendChild(s);
})();
</script>
</section>
</article>
</div>
</main>
<!-- Links to Previous/Next posts -->
<aside class="read-next outer">
<div class="inner">
<div class="read-next-feed">
<article class="read-next-card"
style="background-image: url(/assets/built/images/blog-cover1.png)"
>
<header class="read-next-card-header">
<small class="read-next-card-header-sitetitle">— Darron's Devlog —</small>
<h3 class="read-next-card-header-title"><a href="/tag/projects/">Projects</a></h3>
</header>
<div class="read-next-divider"><svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 24 24"><path d="M13 14.5s2 3 5 3 5.5-2.463 5.5-5.5S21 6.5 18 6.5c-5 0-7 11-12 11C2.962 17.5.5 15.037.5 12S3 6.5 6 6.5s4.5 3.5 4.5 3.5"/></svg>
</div>