Merge branch 'master' of https://github.com/Rintarooo/VRP_MHA

PyTorch/Csv/VRP20_train_0821_00_58.csv

@@ -0,0 +1,319 @@
+time,epoch,batch,loss,cost
+3min55sec,0,0,-1.483,14.204
+0min1sec,0,10,29.913,12.038
+0min1sec,0,20,21.958,11.255
+0min1sec,0,30,17.109,10.689
+0min1sec,0,40,13.896,10.266
+0min1sec,0,50,11.683,9.930
+0min1sec,0,60,9.994,9.660
+0min1sec,0,70,8.711,9.443
+0min1sec,0,80,7.670,9.271
+0min1sec,0,90,6.865,9.129
+0min1sec,0,100,6.240,9.007
+0min1sec,0,110,5.706,8.902
+0min1sec,0,120,5.237,8.812
+0min1sec,0,130,4.851,8.732
+0min1sec,0,140,4.522,8.661
+0min1sec,0,150,4.223,8.598
+0min1sec,0,160,3.956,8.542
+0min1sec,0,170,3.720,8.493
+0min1sec,0,180,3.510,8.447
+0min1sec,0,190,3.329,8.407
+0min1sec,0,200,3.171,8.369
+0min1sec,0,210,3.024,8.334
+0min1sec,0,220,2.888,8.301
+0min1sec,0,230,2.764,8.270
+0min1sec,0,240,2.654,8.241
+0min1sec,0,250,2.544,8.214
+0min1sec,0,260,2.454,8.189
+0min1sec,0,270,2.365,8.165
+0min1sec,0,280,2.278,8.142
+0min1sec,0,290,2.200,8.120
+0min1sec,0,300,2.136,8.098
+0min1sec,0,310,2.060,8.079
+0min1sec,0,320,2.001,8.061
+0min1sec,0,330,1.938,8.044
+0min1sec,0,340,1.882,8.028
+0min1sec,0,350,1.829,8.013
+0min1sec,0,360,1.776,7.998
+0min1sec,0,370,1.730,7.983
+0min1sec,0,380,1.690,7.968
+0min1sec,0,390,1.645,7.953
+0min1sec,0,400,1.595,7.940
+0min1sec,0,410,1.556,7.927
+0min1sec,0,420,1.512,7.916
+0min1sec,0,430,1.476,7.905
+0min1sec,0,440,1.445,7.894
+0min1sec,0,450,1.410,7.884
+0min1sec,0,460,1.380,7.874
+0min1sec,0,470,1.353,7.864
+0min1sec,0,480,1.323,7.854
+0min1sec,0,490,1.292,7.844
+0min1sec,0,500,1.267,7.835
+0min1sec,0,510,1.245,7.825
+0min1sec,0,520,1.216,7.816
+0min1sec,0,530,1.194,7.807
+0min1sec,0,540,1.169,7.799
+0min1sec,0,550,1.146,7.791
+0min1sec,0,560,1.127,7.783
+0min1sec,0,570,1.106,7.775
+0min1sec,0,580,1.092,7.768
+0min1sec,0,590,1.074,7.760
+0min1sec,0,600,1.051,7.754
+0min1sec,0,610,1.037,7.747
+0min1sec,0,620,1.020,7.741
+0min1sec,0,630,1.005,7.734
+0min1sec,0,640,0.987,7.728
+0min1sec,0,650,0.972,7.723
+0min1sec,0,660,0.962,7.717
+0min1sec,0,670,0.946,7.711
+0min1sec,0,680,0.932,7.705
+0min1sec,0,690,0.918,7.700
+0min1sec,0,700,0.903,7.695
+0min1sec,0,710,0.893,7.690
+0min1sec,0,720,0.882,7.685
+0min1sec,0,730,0.873,7.680
+0min1sec,0,740,0.859,7.675
+0min1sec,0,750,0.850,7.670
+0min1sec,0,760,0.838,7.665
+0min1sec,0,770,0.827,7.660
+0min1sec,0,780,0.815,7.656
+0min1sec,0,790,0.805,7.652
+0min1sec,0,800,0.797,7.648
+0min1sec,0,810,0.783,7.644
+0min1sec,0,820,0.776,7.640
+0min1sec,0,830,0.767,7.636
+0min1sec,0,840,0.757,7.632
+0min1sec,0,850,0.748,7.628
+0min1sec,0,860,0.739,7.624
+0min1sec,0,870,0.730,7.621
+0min1sec,0,880,0.722,7.617
+0min1sec,0,890,0.713,7.613
+0min1sec,0,900,0.705,7.609
+0min1sec,0,910,0.698,7.606
+0min1sec,0,920,0.691,7.602
+0min1sec,0,930,0.684,7.599
+0min1sec,0,940,0.679,7.596
+0min1sec,0,950,0.672,7.592
+0min1sec,0,960,0.668,7.589
+0min1sec,0,970,0.663,7.585
+0min1sec,0,980,0.658,7.582
+0min1sec,0,990,0.652,7.578
+0min1sec,0,1000,0.647,7.575
+0min1sec,0,1010,0.640,7.572
+0min1sec,0,1020,0.636,7.568
+0min1sec,0,1030,0.630,7.565
+0min1sec,0,1040,0.624,7.562
+0min1sec,0,1050,0.619,7.559
+0min1sec,0,1060,0.612,7.556
+0min1sec,0,1070,0.609,7.553
+0min1sec,0,1080,0.604,7.550
+0min1sec,0,1090,0.599,7.547
+0min1sec,0,1100,0.595,7.545
+0min1sec,0,1110,0.592,7.542
+0min1sec,0,1120,0.586,7.539
+0min1sec,0,1130,0.583,7.536
+0min1sec,0,1140,0.579,7.534
+0min1sec,0,1150,0.573,7.531
+0min1sec,0,1160,0.568,7.529
+0min1sec,0,1170,0.564,7.527
+0min1sec,0,1180,0.558,7.524
+0min1sec,0,1190,0.555,7.522
+0min1sec,0,1200,0.551,7.520
+0min1sec,0,1210,0.548,7.517
+0min1sec,0,1220,0.546,7.515
+0min1sec,0,1230,0.542,7.512
+0min1sec,0,1240,0.538,7.510
+0min1sec,0,1250,0.537,7.508
+0min1sec,0,1260,0.535,7.505
+0min1sec,0,1270,0.530,7.503
+0min1sec,0,1280,0.527,7.501
+0min1sec,0,1290,0.524,7.499
+0min1sec,0,1300,0.520,7.497
+0min1sec,0,1310,0.517,7.495
+0min1sec,0,1320,0.516,7.493
+0min1sec,0,1330,0.514,7.490
+0min1sec,0,1340,0.511,7.488
+0min1sec,0,1350,0.507,7.486
+0min1sec,0,1360,0.505,7.484
+0min1sec,0,1370,0.501,7.482
+0min1sec,0,1380,0.499,7.480
+0min1sec,0,1390,0.496,7.478
+0min1sec,0,1400,0.493,7.476
+0min1sec,0,1410,0.492,7.474
+0min1sec,0,1420,0.490,7.472
+0min1sec,0,1430,0.488,7.470
+0min1sec,0,1440,0.485,7.468
+0min1sec,0,1450,0.482,7.466
+0min1sec,0,1460,0.479,7.464
+0min1sec,0,1470,0.478,7.462
+0min1sec,0,1480,0.475,7.461
+0min1sec,0,1490,0.475,7.459
+0min1sec,0,1500,0.473,7.457
+0min1sec,0,1510,0.471,7.455
+0min1sec,0,1520,0.469,7.453
+0min1sec,0,1530,0.466,7.452
+0min1sec,0,1540,0.465,7.450
+0min1sec,0,1550,0.463,7.448
+0min1sec,0,1560,0.461,7.447
+0min1sec,0,1570,0.460,7.445
+0min1sec,0,1580,0.458,7.444
+0min1sec,0,1590,0.457,7.442
+0min1sec,0,1600,0.455,7.441
+0min1sec,0,1610,0.455,7.439
+0min1sec,0,1620,0.453,7.438
+0min1sec,0,1630,0.452,7.436
+0min1sec,0,1640,0.451,7.435
+0min1sec,0,1650,0.449,7.433
+0min1sec,0,1660,0.448,7.431
+0min1sec,0,1670,0.445,7.430
+0min1sec,0,1680,0.443,7.429
+0min1sec,0,1690,0.442,7.427
+0min1sec,0,1700,0.440,7.426
+0min1sec,0,1710,0.436,7.424
+0min1sec,0,1720,0.435,7.423
+0min1sec,0,1730,0.433,7.421
+0min1sec,0,1740,0.431,7.420
+0min1sec,0,1750,0.429,7.419
+0min1sec,0,1760,0.429,7.418
+0min1sec,0,1770,0.427,7.416
+0min1sec,0,1780,0.425,7.415
+0min1sec,0,1790,0.424,7.414
+0min1sec,0,1800,0.422,7.412
+0min1sec,0,1810,0.421,7.411
+0min1sec,0,1820,0.420,7.409
+0min1sec,0,1830,0.418,7.408
+0min1sec,0,1840,0.417,7.407
+0min1sec,0,1850,0.414,7.406
+0min1sec,0,1860,0.413,7.405
+0min1sec,0,1870,0.412,7.404
+0min1sec,0,1880,0.411,7.402
+0min1sec,0,1890,0.410,7.401
+0min1sec,0,1900,0.408,7.400
+0min1sec,0,1910,0.407,7.399
+0min1sec,0,1920,0.406,7.397
+0min1sec,0,1930,0.404,7.396
+0min1sec,0,1940,0.403,7.395
+0min1sec,0,1950,0.399,7.394
+0min1sec,0,1960,0.397,7.393
+0min1sec,0,1970,0.397,7.392
+0min1sec,0,1980,0.396,7.391
+0min1sec,0,1990,0.395,7.390
+0min1sec,0,2000,0.394,7.389
+0min1sec,0,2010,0.392,7.388
+0min1sec,0,2020,0.392,7.386
+0min1sec,0,2030,0.391,7.386
+0min1sec,0,2040,0.389,7.385
+0min1sec,0,2050,0.389,7.384
+0min1sec,0,2060,0.386,7.383
+0min1sec,0,2070,0.385,7.382
+0min1sec,0,2080,0.384,7.381
+0min1sec,0,2090,0.383,7.380
+0min1sec,0,2100,0.383,7.379
+0min1sec,0,2110,0.382,7.378
+0min1sec,0,2120,0.381,7.376
+0min1sec,0,2130,0.380,7.376
+0min1sec,0,2140,0.379,7.375
+0min1sec,0,2150,0.378,7.373
+0min1sec,0,2160,0.377,7.372
+0min1sec,0,2170,0.376,7.372
+0min1sec,0,2180,0.375,7.371
+0min1sec,0,2190,0.375,7.369
+0min1sec,0,2200,0.375,7.368
+0min1sec,0,2210,0.374,7.367
+0min1sec,0,2220,0.374,7.366
+0min1sec,0,2230,0.372,7.365
+0min1sec,0,2240,0.371,7.364
+0min1sec,0,2250,0.371,7.363
+0min1sec,0,2260,0.370,7.362
+0min1sec,0,2270,0.369,7.361
+0min1sec,0,2280,0.368,7.360
+0min1sec,0,2290,0.368,7.360
+0min1sec,0,2300,0.367,7.359
+0min1sec,0,2310,0.366,7.358
+0min1sec,0,2320,0.366,7.357
+0min1sec,0,2330,0.366,7.356
+0min1sec,0,2340,0.365,7.354
+0min1sec,0,2350,0.365,7.353
+0min1sec,0,2360,0.364,7.352
+0min1sec,0,2370,0.364,7.351
+0min1sec,0,2380,0.363,7.350
+0min1sec,0,2390,0.363,7.349
+0min1sec,0,2400,0.362,7.348
+0min1sec,0,2410,0.361,7.347
+0min1sec,0,2420,0.361,7.346
+0min1sec,0,2430,0.360,7.346
+0min1sec,0,2440,0.360,7.345
+0min1sec,0,2450,0.360,7.344
+0min1sec,0,2460,0.359,7.343
+0min1sec,0,2470,0.359,7.342
+0min1sec,0,2480,0.358,7.341
+0min1sec,0,2490,0.358,7.340
+5min16sec,1,0,0.352,7.108
+0min1sec,1,10,-0.662,7.134
+0min1sec,1,20,-0.784,7.156
+0min1sec,1,30,-0.771,7.159
+0min1sec,1,40,-0.758,7.157
+0min1sec,1,50,-0.732,7.153
+0min1sec,1,60,-0.734,7.155
+0min1sec,1,70,-0.749,7.158
+0min1sec,1,80,-0.774,7.158
+0min1sec,1,90,-0.798,7.160
+0min1sec,1,100,-0.794,7.158
+0min1sec,1,110,-0.809,7.158
+0min1sec,1,120,-0.798,7.158
+0min1sec,1,130,-0.805,7.159
+0min1sec,1,140,-0.819,7.161
+0min1sec,1,150,-0.790,7.159
+0min1sec,1,160,-0.774,7.158
+0min1sec,1,170,-0.798,7.160
+0min1sec,1,180,-0.800,7.161
+0min1sec,1,190,-0.798,7.161
+0min1sec,1,200,-0.789,7.161
+0min1sec,1,210,-0.798,7.161
+0min1sec,1,220,-0.809,7.162
+0min1sec,1,230,-0.825,7.164
+0min1sec,1,240,-0.824,7.164
+0min1sec,1,250,-0.813,7.164
+0min1sec,1,260,-0.822,7.166
+0min1sec,1,270,-0.822,7.166
+0min1sec,1,280,-0.832,7.168
+0min1sec,1,290,-0.835,7.169
+0min1sec,1,300,-0.841,7.170
+0min1sec,1,310,-0.857,7.171
+0min1sec,1,320,-0.865,7.172
+0min1sec,1,330,-0.869,7.172
+0min1sec,1,340,-0.869,7.172
+0min1sec,1,350,-0.863,7.171
+0min1sec,1,360,-0.862,7.170
+0min1sec,1,370,-0.854,7.169
+0min1sec,1,380,-0.839,7.168
+0min1sec,1,390,-0.820,7.166
+0min1sec,1,400,-0.817,7.165
+0min1sec,1,410,-0.824,7.166
+0min1sec,1,420,-0.821,7.165
+0min1sec,1,430,-0.820,7.165
+0min1sec,1,440,-0.816,7.164
+0min1sec,1,450,-0.803,7.163
+0min1sec,1,460,-0.796,7.163
+0min1sec,1,470,-0.796,7.163
+0min1sec,1,480,-0.795,7.163
+0min1sec,1,490,-0.793,7.162
+0min1sec,1,500,-0.799,7.163
+0min1sec,1,510,-0.795,7.163
+0min1sec,1,520,-0.787,7.163
+0min1sec,1,530,-0.786,7.162
+0min1sec,1,540,-0.780,7.161
+0min1sec,1,550,-0.769,7.160
+0min1sec,1,560,-0.762,7.159
+0min1sec,1,570,-0.759,7.159
+0min1sec,1,580,-0.754,7.158
+0min1sec,1,590,-0.751,7.158
+0min1sec,1,600,-0.752,7.158
+0min1sec,1,610,-0.751,7.157
+0min1sec,1,620,-0.749,7.157
+0min1sec,1,630,-0.752,7.157
+0min1sec,1,640,-0.749,7.158
+0min1sec,1,650,-0.747,7.157
+0min1sec,1,660,-0.745,7.157
+0min1sec,1,670,-0.739,7.156

PyTorch/Csv/param_VRP20_train_0821_00_58.csv

@@ -0,0 +1,24 @@
+mode,train
+seed,123
+n_customer,20
+batch,512
+batch_steps,2500
+batch_verbose,10
+n_rollout_samples,10000
+epochs,20
+embed_dim,128
+n_heads,8
+tanh_clipping,10.0
+n_encode_layers,3
+lr,0.0001
+warmup_beta,0.8
+wp_epochs,1
+islogger,True
+log_dir,./Csv/
+weight_dir,./Weights/
+pkl_dir,./Pkl/
+cuda_dv,0
+task,VRP20_train
+dump_date,0821_00_58
+pkl_path,./Pkl/VRP20_train.pkl
+n_samples,1280000

PyTorch/plot.py

@@ -115,17 +115,17 @@ def plot_route(data, pi, title, cost, idx_in_batch = 0):
 
 if __name__ == '__main__':
 	args = test_parser()
-	t1 = time()
 	pretrained = load_model(args.path, embed_dim = 128, n_customer = args.n_customer, n_encode_layers = 3)
 	data = generate_data(n_samples = 2, n_customer = args.n_customer, seed = args.seed) 
 	device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
 	pretrained = pretrained.to(device)
 	data = list(map(lambda x: x.to(device), data))
 	pretrained.eval()
 	with torch.no_grad():
+		t1 = time()
 		cost, _, pi = pretrained(data, return_pi = True)
 	plot_route(data, pi, 'Pretrained', cost[0], 0)
 	# model = AttentionModel(embed_dim = 128, n_encode_layers = 3, n_heads = 8, tanh_clipping = 10., FF_hidden = 512)
 	# cost, _, pi = model(data, return_pi = True)
 	# plot_route(data, pi, 'Untrained', cost[idx_min], idx_min)
-	print(f'inference time: {time()-t1}s')
+	print(f'inference time: {time()-t1}s')