In [45]:
import pickle
%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
In [33]:
df = pickle.load( open( "2015-12-12-mlpexperiments_results7.p", "rb" ) )
df.shape
Out[33]:
(15, 20)
In [34]:
df2 = pickle.load(open("2015-12-12-mlpexperiments_results8.p", "rb"))
In [58]:
df = df2.append(df)
df.shape
Out[58]:
(45, 23)
In [60]:
df
Out[60]:
B_cm
K_cm
N_cm
P_cm
Q_cm
R_cm
b_cm
black_cm
cm_overall
epochs
...
num_layers
overall_acc
p_cm
pct_white
q_cm
r_cm
test_size
training_size
white_cm
width
1
[[255, 132], [128, 264]]
[[214, 141], [86, 273]]
[[280, 123], [114, 280]]
[[418, 218], [69, 554]]
[[267, 149], [114, 356]]
[[296, 182], [187, 337]]
[[248, 58], [207, 95]]
[[1799, 498], [1345, 921]]
[[3529, 2043], [1443, 2985]]
5
...
NaN
NaN
[[514, 89], [325, 245]]
0.5028
[[298, 70], [196, 153]]
[[290, 137], [252, 193]]
10000
100000
[[1730, 945], [698, 2064]]
NaN
1
[[254, 133], [107, 285]]
[[201, 154], [73, 286]]
[[272, 131], [111, 283]]
[[414, 222], [71, 552]]
[[261, 155], [99, 371]]
[[283, 195], [146, 378]]
[[109, 197], [11, 291]]
[[800, 1497], [165, 2101]]
[[2485, 772], [2487, 4256]]
5
...
NaN
NaN
[[212, 391], [11, 559]]
0.5028
[[141, 227], [21, 328]]
[[137, 290], [60, 385]]
10000
100000
[[1685, 990], [607, 2155]]
NaN
1
[[163, 224], [37, 355]]
[[174, 181], [71, 288]]
[[177, 226], [27, 367]]
[[266, 370], [33, 590]]
[[171, 245], [45, 425]]
[[225, 253], [91, 433]]
[[230, 76], [180, 122]]
[[1743, 554], [1049, 1217]]
[[2919, 1353], [2053, 3675]]
5
...
NaN
NaN
[[512, 91], [128, 442]]
0.5028
[[283, 85], [170, 179]]
[[277, 150], [268, 177]]
10000
100000
[[1176, 1499], [304, 2458]]
NaN
1
[[224, 163], [88, 304]]
[[206, 149], [86, 273]]
[[249, 154], [64, 330]]
[[376, 260], [62, 561]]
[[246, 170], [86, 384]]
[[274, 204], [134, 390]]
[[200, 106], [100, 202]]
[[1504, 793], [635, 1631]]
[[3079, 1155], [1893, 3873]]
5
...
NaN
NaN
[[446, 157], [51, 519]]
0.5028
[[241, 127], [93, 256]]
[[237, 190], [170, 275]]
10000
100000
[[1575, 1100], [520, 2242]]
NaN
1
[[313, 74], [253, 139]]
[[240, 115], [143, 216]]
[[336, 67], [211, 183]]
[[530, 106], [207, 416]]
[[322, 94], [218, 252]]
[[350, 128], [273, 251]]
[[150, 156], [46, 256]]
[[1131, 1166], [404, 1862]]
[[3222, 1709], [1750, 3319]]
5
...
NaN
NaN
[[297, 306], [24, 546]]
0.5028
[[182, 186], [55, 294]]
[[211, 216], [126, 319]]
10000
100000
[[2091, 584], [1305, 1457]]
NaN
1
[[223, 164], [74, 318]]
[[186, 169], [72, 287]]
[[255, 148], [80, 314]]
[[375, 261], [52, 571]]
[[239, 177], [70, 400]]
[[263, 215], [104, 420]]
[[167, 139], [48, 254]]
[[1214, 1083], [388, 1878]]
[[2755, 840], [2217, 4188]]
5
...
NaN
NaN
[[350, 253], [25, 545]]
0.5028
[[193, 175], [65, 284]]
[[201, 226], [121, 324]]
10000
100000
[[1541, 1134], [452, 2310]]
NaN
1
[[291, 96], [196, 196]]
[[204, 151], [97, 262]]
[[316, 87], [169, 225]]
[[495, 141], [137, 486]]
[[306, 110], [170, 300]]
[[306, 172], [194, 330]]
[[136, 170], [24, 278]]
[[958, 1339], [180, 2086]]
[[2876, 1143], [2096, 3885]]
5
...
NaN
NaN
[[283, 320], [14, 556]]
0.5028
[[158, 210], [18, 331]]
[[147, 280], [54, 391]]
10000
100000
[[1918, 757], [963, 1799]]
NaN
1
[[167, 220], [24, 368]]
[[174, 181], [47, 312]]
[[204, 199], [22, 372]]
[[281, 355], [40, 583]]
[[188, 228], [33, 437]]
[[230, 248], [76, 448]]
[[221, 85], [156, 146]]
[[1617, 680], [853, 1413]]
[[2861, 1095], [2111, 3933]]
5
...
NaN
NaN
[[474, 129], [114, 456]]
0.5028
[[273, 95], [131, 218]]
[[248, 179], [193, 252]]
10000
100000
[[1244, 1431], [242, 2520]]
NaN
1
[[282, 105], [156, 236]]
[[209, 146], [87, 272]]
[[293, 110], [123, 271]]
[[448, 188], [88, 535]]
[[278, 138], [120, 350]]
[[291, 187], [177, 347]]
[[180, 126], [84, 218]]
[[1328, 969], [507, 1759]]
[[3129, 1258], [1843, 3770]]
5
...
NaN
NaN
[[382, 221], [38, 532]]
0.5028
[[222, 146], [67, 282]]
[[207, 220], [146, 299]]
10000
100000
[[1801, 874], [751, 2011]]
NaN
1
[[279, 108], [173, 219]]
[[202, 153], [78, 281]]
[[307, 96], [154, 240]]
[[464, 172], [116, 507]]
[[291, 125], [139, 331]]
[[297, 181], [168, 356]]
[[214, 92], [133, 169]]
[[1536, 761], [745, 1521]]
[[3376, 1573], [1596, 3455]]
5
...
NaN
NaN
[[465, 138], [89, 481]]
0.5028
[[250, 118], [114, 235]]
[[226, 201], [173, 272]]
10000
100000
[[1840, 835], [828, 1934]]
NaN
1
[[180, 207], [27, 365]]
[[135, 220], [29, 330]]
[[201, 202], [22, 372]]
[[274, 362], [28, 595]]
[[181, 235], [19, 451]]
[[197, 281], [46, 478]]
[[149, 157], [26, 276]]
[[1071, 1226], [190, 2076]]
[[2239, 361], [2733, 4667]]
5
...
NaN
NaN
[[343, 260], [20, 550]]
0.5028
[[171, 197], [25, 324]]
[[149, 278], [45, 400]]
10000
100000
[[1168, 1507], [171, 2591]]
NaN
1
[[142, 245], [15, 377]]
[[131, 224], [23, 336]]
[[150, 253], [12, 382]]
[[229, 407], [25, 598]]
[[162, 254], [18, 452]]
[[187, 291], [42, 482]]
[[230, 76], [153, 149]]
[[1607, 690], [953, 1313]]
[[2608, 1088], [2364, 3940]]
5
...
NaN
NaN
[[482, 121], [216, 354]]
0.5028
[[264, 104], [142, 207]]
[[237, 190], [171, 274]]
10000
100000
[[1001, 1674], [135, 2627]]
NaN
1
[[305, 82], [211, 181]]
[[223, 132], [121, 238]]
[[326, 77], [175, 219]]
[[501, 135], [123, 500]]
[[327, 89], [186, 284]]
[[337, 141], [241, 283]]
[[181, 125], [69, 233]]
[[1321, 976], [473, 1793]]
[[3340, 1530], [1632, 3498]]
5
...
NaN
NaN
[[380, 223], [35, 535]]
0.5028
[[214, 154], [79, 270]]
[[218, 209], [140, 305]]
10000
100000
[[2019, 656], [1057, 1705]]
NaN
1
[[230, 157], [78, 314]]
[[195, 160], [62, 297]]
[[256, 147], [87, 307]]
[[388, 248], [47, 576]]
[[241, 175], [72, 398]]
[[274, 204], [122, 402]]
[[131, 175], [30, 272]]
[[986, 1311], [252, 2014]]
[[2570, 720], [2402, 4308]]
5
...
NaN
NaN
[[272, 331], [22, 548]]
0.5028
[[173, 195], [38, 311]]
[[162, 265], [80, 365]]
10000
100000
[[1584, 1091], [468, 2294]]
NaN
1
[[266, 121], [118, 274]]
[[188, 167], [62, 297]]
[[285, 118], [100, 294]]
[[426, 210], [66, 557]]
[[279, 137], [98, 372]]
[[283, 195], [157, 367]]
[[182, 124], [77, 225]]
[[1293, 1004], [418, 1848]]
[[3020, 1019], [1952, 4009]]
5
...
NaN
NaN
[[398, 205], [38, 532]]
0.5028
[[209, 159], [53, 296]]
[[175, 252], [95, 350]]
10000
100000
[[1727, 948], [601, 2161]]
NaN
1
[[277, 110], [159, 233]]
[[231, 124], [100, 259]]
[[295, 108], [138, 256]]
[[463, 173], [96, 527]]
[[284, 132], [135, 335]]
[[311, 167], [187, 337]]
[[146, 160], [39, 263]]
[[1112, 1185], [335, 1931]]
[[2973, 1150], [1999, 3878]]
5
...
3
0.6851
[[304, 299], [18, 552]]
0.5028
[[190, 178], [45, 304]]
[[190, 237], [113, 332]]
10000
100000
[[1861, 814], [815, 1947]]
100
1
[[328, 59], [262, 130]]
[[243, 112], [155, 204]]
[[345, 58], [236, 158]]
[[541, 95], [268, 355]]
[[345, 71], [246, 224]]
[[367, 111], [284, 240]]
[[244, 62], [196, 106]]
[[1757, 540], [1210, 1056]]
[[3926, 2661], [1046, 2367]]
5
...
3
0.6293
[[506, 97], [278, 292]]
0.5028
[[290, 78], [170, 179]]
[[284, 143], [229, 216]]
10000
100000
[[2169, 506], [1451, 1311]]
200
1
[[269, 118], [114, 278]]
[[224, 131], [103, 256]]
[[279, 124], [94, 300]]
[[426, 210], [77, 546]]
[[270, 146], [127, 343]]
[[327, 151], [199, 325]]
[[246, 60], [214, 88]]
[[1852, 445], [1175, 1091]]
[[3647, 1889], [1325, 3139]]
5
...
3
0.6786
[[517, 86], [128, 442]]
0.5028
[[305, 63], [177, 172]]
[[318, 109], [301, 144]]
10000
100000
[[1795, 880], [714, 2048]]
300
1
[[331, 56], [279, 113]]
[[258, 97], [179, 180]]
[[351, 52], [240, 154]]
[[555, 81], [246, 377]]
[[352, 64], [293, 177]]
[[380, 98], [339, 185]]
[[167, 139], [56, 246]]
[[1275, 1022], [438, 1828]]
[[3502, 2014], [1470, 3014]]
5
...
3
0.6516
[[364, 239], [40, 530]]
0.5028
[[202, 166], [60, 289]]
[[208, 219], [134, 311]]
10000
100000
[[2227, 448], [1576, 1186]]
400
1
[[261, 126], [97, 295]]
[[215, 140], [92, 267]]
[[269, 134], [81, 313]]
[[415, 221], [67, 556]]
[[258, 158], [90, 380]]
[[309, 169], [178, 346]]
[[186, 120], [73, 229]]
[[1371, 926], [522, 1744]]
[[3098, 1127], [1874, 3901]]
5
...
4
0.6999
[[404, 199], [47, 523]]
0.5028
[[212, 156], [70, 279]]
[[228, 199], [156, 289]]
10000
100000
[[1727, 948], [605, 2157]]
100
1
[[231, 156], [69, 323]]
[[151, 204], [20, 339]]
[[239, 164], [67, 327]]
[[377, 259], [38, 585]]
[[222, 194], [45, 425]]
[[215, 263], [61, 463]]
[[111, 195], [5, 297]]
[[795, 1502], [72, 2194]]
[[2230, 372], [2742, 4656]]
5
...
4
0.6886
[[238, 365], [13, 557]]
0.5028
[[139, 229], [9, 340]]
[[121, 306], [21, 424]]
10000
100000
[[1435, 1240], [300, 2462]]
200
1
[[242, 145], [58, 334]]
[[167, 188], [29, 330]]
[[260, 143], [66, 328]]
[[391, 245], [51, 572]]
[[248, 168], [46, 424]]
[[243, 235], [84, 440]]
[[174, 132], [46, 256]]
[[1238, 1059], [286, 1980]]
[[2789, 620], [2183, 4408]]
5
...
4
0.7197
[[382, 221], [27, 543]]
0.5028
[[208, 160], [47, 302]]
[[174, 253], [64, 381]]
10000
100000
[[1551, 1124], [334, 2428]]
300
1
[[302, 85], [217, 175]]
[[211, 144], [101, 258]]
[[320, 83], [191, 203]]
[[500, 136], [150, 473]]
[[321, 95], [194, 276]]
[[320, 158], [232, 292]]
[[198, 108], [77, 225]]
[[1444, 853], [500, 1766]]
[[3418, 1585], [1554, 3443]]
5
...
4
0.6861
[[421, 182], [42, 528]]
0.5028
[[234, 134], [83, 266]]
[[232, 195], [141, 304]]
10000
100000
[[1974, 701], [1085, 1677]]
400
1
[[245, 142], [83, 309]]
[[171, 184], [47, 312]]
[[279, 124], [87, 307]]
[[410, 226], [57, 566]]
[[266, 150], [82, 388]]
[[282, 196], [135, 389]]
[[115, 191], [15, 287]]
[[879, 1418], [146, 2120]]
[[2532, 637], [2440, 4391]]
5
...
5
0.6923
[[254, 349], [14, 556]]
0.5028
[[158, 210], [15, 334]]
[[135, 292], [44, 401]]
10000
100000
[[1653, 1022], [491, 2271]]
100
1
[[276, 111], [129, 263]]
[[179, 176], [52, 307]]
[[297, 106], [114, 280]]
[[456, 180], [88, 535]]
[[292, 124], [131, 339]]
[[283, 195], [145, 379]]
[[114, 192], [9, 293]]
[[841, 1456], [81, 2185]]
[[2624, 740], [2348, 4288]]
5
...
5
0.6912
[[260, 343], [10, 560]]
0.5028
[[146, 222], [12, 337]]
[[115, 312], [23, 422]]
10000
100000
[[1783, 892], [659, 2103]]
200
1
[[199, 188], [41, 351]]
[[146, 209], [10, 349]]
[[209, 194], [43, 351]]
[[324, 312], [29, 594]]
[[196, 220], [24, 446]]
[[198, 280], [27, 497]]
[[165, 141], [40, 262]]
[[1149, 1148], [242, 2024]]
[[2421, 416], [2551, 4612]]
5
...
5
0.7033
[[341, 262], [30, 540]]
0.5028
[[197, 171], [37, 312]]
[[161, 266], [58, 387]]
10000
100000
[[1272, 1403], [174, 2588]]
300
1
[[373, 14], [351, 41]]
[[338, 17], [298, 61]]
[[389, 14], [335, 59]]
[[611, 25], [439, 184]]
[[398, 18], [408, 62]]
[[460, 18], [465, 59]]
[[269, 37], [212, 90]]
[[2030, 267], [1354, 912]]
[[4599, 3650], [373, 1378]]
5
...
5
0.5977
[[532, 71], [162, 408]]
0.5028
[[326, 42], [216, 133]]
[[381, 46], [356, 89]]
10000
100000
[[2569, 106], [2296, 466]]
400
1
[[245, 142], [72, 320]]
[[158, 197], [27, 332]]
[[267, 136], [79, 315]]
[[396, 240], [46, 577]]
[[247, 169], [68, 402]]
[[241, 237], [90, 434]]
[[100, 206], [4, 298]]
[[752, 1545], [49, 2217]]
[[2306, 431], [2666, 4597]]
5
...
6
0.6903
[[230, 373], [5, 565]]
0.5028
[[128, 240], [2, 347]]
[[101, 326], [18, 427]]
10000
100000
[[1554, 1121], [382, 2380]]
100
1
[[197, 190], [32, 360]]
[[138, 217], [9, 350]]
[[220, 183], [18, 376]]
[[324, 312], [29, 594]]
[[216, 200], [21, 449]]
[[214, 264], [39, 485]]
[[133, 173], [23, 279]]
[[961, 1336], [117, 2149]]
[[2270, 265], [2702, 4763]]
5
...
6
0.7033
[[315, 288], [11, 559]]
0.5028
[[159, 209], [15, 334]]
[[113, 314], [18, 427]]
10000
100000
[[1309, 1366], [148, 2614]]
200
1
[[162, 225], [13, 379]]
[[131, 224], [6, 353]]
[[165, 238], [5, 389]]
[[244, 392], [21, 602]]
[[165, 251], [14, 456]]
[[187, 291], [21, 503]]
[[175, 131], [43, 259]]
[[1190, 1107], [263, 2003]]
[[2244, 343], [2728, 4685]]
5
...
6
0.6929
[[368, 235], [27, 543]]
0.5028
[[194, 174], [41, 308]]
[[168, 259], [54, 391]]
10000
100000
[[1054, 1621], [80, 2682]]
300
1
[[255, 132], [128, 264]]
[[214, 141], [86, 273]]
[[280, 123], [114, 280]]
[[418, 218], [69, 554]]
[[267, 149], [114, 356]]
[[296, 182], [187, 337]]
[[248, 58], [207, 95]]
[[1799, 498], [1345, 921]]
[[3529, 2043], [1443, 2985]]
5
...
3
0.6514
[[514, 89], [325, 245]]
0.5028
[[298, 70], [196, 153]]
[[290, 137], [252, 193]]
10000
100000
[[1730, 945], [698, 2064]]
5
1
[[254, 133], [107, 285]]
[[201, 154], [73, 286]]
[[272, 131], [111, 283]]
[[414, 222], [71, 552]]
[[261, 155], [99, 371]]
[[283, 195], [146, 378]]
[[109, 197], [11, 291]]
[[800, 1497], [165, 2101]]
[[2485, 772], [2487, 4256]]
5
...
3
0.6741
[[212, 391], [11, 559]]
0.5028
[[141, 227], [21, 328]]
[[137, 290], [60, 385]]
10000
100000
[[1685, 990], [607, 2155]]
25
1
[[163, 224], [37, 355]]
[[174, 181], [71, 288]]
[[177, 226], [27, 367]]
[[266, 370], [33, 590]]
[[171, 245], [45, 425]]
[[225, 253], [91, 433]]
[[230, 76], [180, 122]]
[[1743, 554], [1049, 1217]]
[[2919, 1353], [2053, 3675]]
5
...
3
0.6594
[[512, 91], [128, 442]]
0.5028
[[283, 85], [170, 179]]
[[277, 150], [268, 177]]
10000
100000
[[1176, 1499], [304, 2458]]
50
1
[[224, 163], [88, 304]]
[[206, 149], [86, 273]]
[[249, 154], [64, 330]]
[[376, 260], [62, 561]]
[[246, 170], [86, 384]]
[[274, 204], [134, 390]]
[[200, 106], [100, 202]]
[[1504, 793], [635, 1631]]
[[3079, 1155], [1893, 3873]]
5
...
3
0.6952
[[446, 157], [51, 519]]
0.5028
[[241, 127], [93, 256]]
[[237, 190], [170, 275]]
10000
100000
[[1575, 1100], [520, 2242]]
75
1
[[313, 74], [253, 139]]
[[240, 115], [143, 216]]
[[336, 67], [211, 183]]
[[530, 106], [207, 416]]
[[322, 94], [218, 252]]
[[350, 128], [273, 251]]
[[150, 156], [46, 256]]
[[1131, 1166], [404, 1862]]
[[3222, 1709], [1750, 3319]]
5
...
4
0.6541
[[297, 306], [24, 546]]
0.5028
[[182, 186], [55, 294]]
[[211, 216], [126, 319]]
10000
100000
[[2091, 584], [1305, 1457]]
5
1
[[223, 164], [74, 318]]
[[186, 169], [72, 287]]
[[255, 148], [80, 314]]
[[375, 261], [52, 571]]
[[239, 177], [70, 400]]
[[263, 215], [104, 420]]
[[167, 139], [48, 254]]
[[1214, 1083], [388, 1878]]
[[2755, 840], [2217, 4188]]
5
...
4
0.6943
[[350, 253], [25, 545]]
0.5028
[[193, 175], [65, 284]]
[[201, 226], [121, 324]]
10000
100000
[[1541, 1134], [452, 2310]]
25
1
[[291, 96], [196, 196]]
[[204, 151], [97, 262]]
[[316, 87], [169, 225]]
[[495, 141], [137, 486]]
[[306, 110], [170, 300]]
[[306, 172], [194, 330]]
[[136, 170], [24, 278]]
[[958, 1339], [180, 2086]]
[[2876, 1143], [2096, 3885]]
5
...
4
0.6761
[[283, 320], [14, 556]]
0.5028
[[158, 210], [18, 331]]
[[147, 280], [54, 391]]
10000
100000
[[1918, 757], [963, 1799]]
50
1
[[167, 220], [24, 368]]
[[174, 181], [47, 312]]
[[204, 199], [22, 372]]
[[281, 355], [40, 583]]
[[188, 228], [33, 437]]
[[230, 248], [76, 448]]
[[221, 85], [156, 146]]
[[1617, 680], [853, 1413]]
[[2861, 1095], [2111, 3933]]
5
...
4
0.6794
[[474, 129], [114, 456]]
0.5028
[[273, 95], [131, 218]]
[[248, 179], [193, 252]]
10000
100000
[[1244, 1431], [242, 2520]]
75
1
[[282, 105], [156, 236]]
[[209, 146], [87, 272]]
[[293, 110], [123, 271]]
[[448, 188], [88, 535]]
[[278, 138], [120, 350]]
[[291, 187], [177, 347]]
[[180, 126], [84, 218]]
[[1328, 969], [507, 1759]]
[[3129, 1258], [1843, 3770]]
5
...
5
0.6899
[[382, 221], [38, 532]]
0.5028
[[222, 146], [67, 282]]
[[207, 220], [146, 299]]
10000
100000
[[1801, 874], [751, 2011]]
5
1
[[279, 108], [173, 219]]
[[202, 153], [78, 281]]
[[307, 96], [154, 240]]
[[464, 172], [116, 507]]
[[291, 125], [139, 331]]
[[297, 181], [168, 356]]
[[214, 92], [133, 169]]
[[1536, 761], [745, 1521]]
[[3376, 1573], [1596, 3455]]
5
...
5
0.6831
[[465, 138], [89, 481]]
0.5028
[[250, 118], [114, 235]]
[[226, 201], [173, 272]]
10000
100000
[[1840, 835], [828, 1934]]
25
1
[[180, 207], [27, 365]]
[[135, 220], [29, 330]]
[[201, 202], [22, 372]]
[[274, 362], [28, 595]]
[[181, 235], [19, 451]]
[[197, 281], [46, 478]]
[[149, 157], [26, 276]]
[[1071, 1226], [190, 2076]]
[[2239, 361], [2733, 4667]]
5
...
5
0.6906
[[343, 260], [20, 550]]
0.5028
[[171, 197], [25, 324]]
[[149, 278], [45, 400]]
10000
100000
[[1168, 1507], [171, 2591]]
50
1
[[142, 245], [15, 377]]
[[131, 224], [23, 336]]
[[150, 253], [12, 382]]
[[229, 407], [25, 598]]
[[162, 254], [18, 452]]
[[187, 291], [42, 482]]
[[230, 76], [153, 149]]
[[1607, 690], [953, 1313]]
[[2608, 1088], [2364, 3940]]
5
...
5
0.6548
[[482, 121], [216, 354]]
0.5028
[[264, 104], [142, 207]]
[[237, 190], [171, 274]]
10000
100000
[[1001, 1674], [135, 2627]]
75
1
[[305, 82], [211, 181]]
[[223, 132], [121, 238]]
[[326, 77], [175, 219]]
[[501, 135], [123, 500]]
[[327, 89], [186, 284]]
[[337, 141], [241, 283]]
[[181, 125], [69, 233]]
[[1321, 976], [473, 1793]]
[[3340, 1530], [1632, 3498]]
5
...
6
0.6838
[[380, 223], [35, 535]]
0.5028
[[214, 154], [79, 270]]
[[218, 209], [140, 305]]
10000
100000
[[2019, 656], [1057, 1705]]
5
1
[[230, 157], [78, 314]]
[[195, 160], [62, 297]]
[[256, 147], [87, 307]]
[[388, 248], [47, 576]]
[[241, 175], [72, 398]]
[[274, 204], [122, 402]]
[[131, 175], [30, 272]]
[[986, 1311], [252, 2014]]
[[2570, 720], [2402, 4308]]
5
...
6
0.6878
[[272, 331], [22, 548]]
0.5028
[[173, 195], [38, 311]]
[[162, 265], [80, 365]]
10000
100000
[[1584, 1091], [468, 2294]]
25
1
[[266, 121], [118, 274]]
[[188, 167], [62, 297]]
[[285, 118], [100, 294]]
[[426, 210], [66, 557]]
[[279, 137], [98, 372]]
[[283, 195], [157, 367]]
[[182, 124], [77, 225]]
[[1293, 1004], [418, 1848]]
[[3020, 1019], [1952, 4009]]
5
...
6
0.7029
[[398, 205], [38, 532]]
0.5028
[[209, 159], [53, 296]]
[[175, 252], [95, 350]]
10000
100000
[[1727, 948], [601, 2161]]
50
45 rows × 23 columns
In [61]:
#methods to decode serialized network json
import json
def get_num_layers(json_str):
# number of actual layers - 5 for input and output / 3 for each hidden + 2 for input and output
return (len(json.loads(json_str)['layers']) - 5) / 3 + 2
def get_first_activation(json_str):
return json.loads(json_str)['layers'][1]['activation']
def get_first_width(json_str):
return json.loads(json_str)['layers'][0]['output_dim']
In [62]:
#convert confusion matrices to accuracy
def cm2accuracy(cm):
return (cm[0][0] + cm[1][1] * 1.0) / sum([sum(a) for a in cm])
df['overall_acc'] = df['cm_overall'].apply(cm2accuracy)
df['width'] = df['network'].apply(get_first_width)
df['num_layers'] = df['network'].apply(get_num_layers)
In [63]:
df
Out[63]:
B_cm
K_cm
N_cm
P_cm
Q_cm
R_cm
b_cm
black_cm
cm_overall
epochs
...
num_layers
overall_acc
p_cm
pct_white
q_cm
r_cm
test_size
training_size
white_cm
width
1
[[255, 132], [128, 264]]
[[214, 141], [86, 273]]
[[280, 123], [114, 280]]
[[418, 218], [69, 554]]
[[267, 149], [114, 356]]
[[296, 182], [187, 337]]
[[248, 58], [207, 95]]
[[1799, 498], [1345, 921]]
[[3529, 2043], [1443, 2985]]
5
...
3
0.6514
[[514, 89], [325, 245]]
0.5028
[[298, 70], [196, 153]]
[[290, 137], [252, 193]]
10000
100000
[[1730, 945], [698, 2064]]
5
1
[[254, 133], [107, 285]]
[[201, 154], [73, 286]]
[[272, 131], [111, 283]]
[[414, 222], [71, 552]]
[[261, 155], [99, 371]]
[[283, 195], [146, 378]]
[[109, 197], [11, 291]]
[[800, 1497], [165, 2101]]
[[2485, 772], [2487, 4256]]
5
...
3
0.6741
[[212, 391], [11, 559]]
0.5028
[[141, 227], [21, 328]]
[[137, 290], [60, 385]]
10000
100000
[[1685, 990], [607, 2155]]
25
1
[[163, 224], [37, 355]]
[[174, 181], [71, 288]]
[[177, 226], [27, 367]]
[[266, 370], [33, 590]]
[[171, 245], [45, 425]]
[[225, 253], [91, 433]]
[[230, 76], [180, 122]]
[[1743, 554], [1049, 1217]]
[[2919, 1353], [2053, 3675]]
5
...
3
0.6594
[[512, 91], [128, 442]]
0.5028
[[283, 85], [170, 179]]
[[277, 150], [268, 177]]
10000
100000
[[1176, 1499], [304, 2458]]
50
1
[[224, 163], [88, 304]]
[[206, 149], [86, 273]]
[[249, 154], [64, 330]]
[[376, 260], [62, 561]]
[[246, 170], [86, 384]]
[[274, 204], [134, 390]]
[[200, 106], [100, 202]]
[[1504, 793], [635, 1631]]
[[3079, 1155], [1893, 3873]]
5
...
3
0.6952
[[446, 157], [51, 519]]
0.5028
[[241, 127], [93, 256]]
[[237, 190], [170, 275]]
10000
100000
[[1575, 1100], [520, 2242]]
75
1
[[313, 74], [253, 139]]
[[240, 115], [143, 216]]
[[336, 67], [211, 183]]
[[530, 106], [207, 416]]
[[322, 94], [218, 252]]
[[350, 128], [273, 251]]
[[150, 156], [46, 256]]
[[1131, 1166], [404, 1862]]
[[3222, 1709], [1750, 3319]]
5
...
4
0.6541
[[297, 306], [24, 546]]
0.5028
[[182, 186], [55, 294]]
[[211, 216], [126, 319]]
10000
100000
[[2091, 584], [1305, 1457]]
5
1
[[223, 164], [74, 318]]
[[186, 169], [72, 287]]
[[255, 148], [80, 314]]
[[375, 261], [52, 571]]
[[239, 177], [70, 400]]
[[263, 215], [104, 420]]
[[167, 139], [48, 254]]
[[1214, 1083], [388, 1878]]
[[2755, 840], [2217, 4188]]
5
...
4
0.6943
[[350, 253], [25, 545]]
0.5028
[[193, 175], [65, 284]]
[[201, 226], [121, 324]]
10000
100000
[[1541, 1134], [452, 2310]]
25
1
[[291, 96], [196, 196]]
[[204, 151], [97, 262]]
[[316, 87], [169, 225]]
[[495, 141], [137, 486]]
[[306, 110], [170, 300]]
[[306, 172], [194, 330]]
[[136, 170], [24, 278]]
[[958, 1339], [180, 2086]]
[[2876, 1143], [2096, 3885]]
5
...
4
0.6761
[[283, 320], [14, 556]]
0.5028
[[158, 210], [18, 331]]
[[147, 280], [54, 391]]
10000
100000
[[1918, 757], [963, 1799]]
50
1
[[167, 220], [24, 368]]
[[174, 181], [47, 312]]
[[204, 199], [22, 372]]
[[281, 355], [40, 583]]
[[188, 228], [33, 437]]
[[230, 248], [76, 448]]
[[221, 85], [156, 146]]
[[1617, 680], [853, 1413]]
[[2861, 1095], [2111, 3933]]
5
...
4
0.6794
[[474, 129], [114, 456]]
0.5028
[[273, 95], [131, 218]]
[[248, 179], [193, 252]]
10000
100000
[[1244, 1431], [242, 2520]]
75
1
[[282, 105], [156, 236]]
[[209, 146], [87, 272]]
[[293, 110], [123, 271]]
[[448, 188], [88, 535]]
[[278, 138], [120, 350]]
[[291, 187], [177, 347]]
[[180, 126], [84, 218]]
[[1328, 969], [507, 1759]]
[[3129, 1258], [1843, 3770]]
5
...
5
0.6899
[[382, 221], [38, 532]]
0.5028
[[222, 146], [67, 282]]
[[207, 220], [146, 299]]
10000
100000
[[1801, 874], [751, 2011]]
5
1
[[279, 108], [173, 219]]
[[202, 153], [78, 281]]
[[307, 96], [154, 240]]
[[464, 172], [116, 507]]
[[291, 125], [139, 331]]
[[297, 181], [168, 356]]
[[214, 92], [133, 169]]
[[1536, 761], [745, 1521]]
[[3376, 1573], [1596, 3455]]
5
...
5
0.6831
[[465, 138], [89, 481]]
0.5028
[[250, 118], [114, 235]]
[[226, 201], [173, 272]]
10000
100000
[[1840, 835], [828, 1934]]
25
1
[[180, 207], [27, 365]]
[[135, 220], [29, 330]]
[[201, 202], [22, 372]]
[[274, 362], [28, 595]]
[[181, 235], [19, 451]]
[[197, 281], [46, 478]]
[[149, 157], [26, 276]]
[[1071, 1226], [190, 2076]]
[[2239, 361], [2733, 4667]]
5
...
5
0.6906
[[343, 260], [20, 550]]
0.5028
[[171, 197], [25, 324]]
[[149, 278], [45, 400]]
10000
100000
[[1168, 1507], [171, 2591]]
50
1
[[142, 245], [15, 377]]
[[131, 224], [23, 336]]
[[150, 253], [12, 382]]
[[229, 407], [25, 598]]
[[162, 254], [18, 452]]
[[187, 291], [42, 482]]
[[230, 76], [153, 149]]
[[1607, 690], [953, 1313]]
[[2608, 1088], [2364, 3940]]
5
...
5
0.6548
[[482, 121], [216, 354]]
0.5028
[[264, 104], [142, 207]]
[[237, 190], [171, 274]]
10000
100000
[[1001, 1674], [135, 2627]]
75
1
[[305, 82], [211, 181]]
[[223, 132], [121, 238]]
[[326, 77], [175, 219]]
[[501, 135], [123, 500]]
[[327, 89], [186, 284]]
[[337, 141], [241, 283]]
[[181, 125], [69, 233]]
[[1321, 976], [473, 1793]]
[[3340, 1530], [1632, 3498]]
5
...
6
0.6838
[[380, 223], [35, 535]]
0.5028
[[214, 154], [79, 270]]
[[218, 209], [140, 305]]
10000
100000
[[2019, 656], [1057, 1705]]
5
1
[[230, 157], [78, 314]]
[[195, 160], [62, 297]]
[[256, 147], [87, 307]]
[[388, 248], [47, 576]]
[[241, 175], [72, 398]]
[[274, 204], [122, 402]]
[[131, 175], [30, 272]]
[[986, 1311], [252, 2014]]
[[2570, 720], [2402, 4308]]
5
...
6
0.6878
[[272, 331], [22, 548]]
0.5028
[[173, 195], [38, 311]]
[[162, 265], [80, 365]]
10000
100000
[[1584, 1091], [468, 2294]]
25
1
[[266, 121], [118, 274]]
[[188, 167], [62, 297]]
[[285, 118], [100, 294]]
[[426, 210], [66, 557]]
[[279, 137], [98, 372]]
[[283, 195], [157, 367]]
[[182, 124], [77, 225]]
[[1293, 1004], [418, 1848]]
[[3020, 1019], [1952, 4009]]
5
...
6
0.7029
[[398, 205], [38, 532]]
0.5028
[[209, 159], [53, 296]]
[[175, 252], [95, 350]]
10000
100000
[[1727, 948], [601, 2161]]
50
1
[[277, 110], [159, 233]]
[[231, 124], [100, 259]]
[[295, 108], [138, 256]]
[[463, 173], [96, 527]]
[[284, 132], [135, 335]]
[[311, 167], [187, 337]]
[[146, 160], [39, 263]]
[[1112, 1185], [335, 1931]]
[[2973, 1150], [1999, 3878]]
5
...
3
0.6851
[[304, 299], [18, 552]]
0.5028
[[190, 178], [45, 304]]
[[190, 237], [113, 332]]
10000
100000
[[1861, 814], [815, 1947]]
100
1
[[328, 59], [262, 130]]
[[243, 112], [155, 204]]
[[345, 58], [236, 158]]
[[541, 95], [268, 355]]
[[345, 71], [246, 224]]
[[367, 111], [284, 240]]
[[244, 62], [196, 106]]
[[1757, 540], [1210, 1056]]
[[3926, 2661], [1046, 2367]]
5
...
3
0.6293
[[506, 97], [278, 292]]
0.5028
[[290, 78], [170, 179]]
[[284, 143], [229, 216]]
10000
100000
[[2169, 506], [1451, 1311]]
200
1
[[269, 118], [114, 278]]
[[224, 131], [103, 256]]
[[279, 124], [94, 300]]
[[426, 210], [77, 546]]
[[270, 146], [127, 343]]
[[327, 151], [199, 325]]
[[246, 60], [214, 88]]
[[1852, 445], [1175, 1091]]
[[3647, 1889], [1325, 3139]]
5
...
3
0.6786
[[517, 86], [128, 442]]
0.5028
[[305, 63], [177, 172]]
[[318, 109], [301, 144]]
10000
100000
[[1795, 880], [714, 2048]]
300
1
[[331, 56], [279, 113]]
[[258, 97], [179, 180]]
[[351, 52], [240, 154]]
[[555, 81], [246, 377]]
[[352, 64], [293, 177]]
[[380, 98], [339, 185]]
[[167, 139], [56, 246]]
[[1275, 1022], [438, 1828]]
[[3502, 2014], [1470, 3014]]
5
...
3
0.6516
[[364, 239], [40, 530]]
0.5028
[[202, 166], [60, 289]]
[[208, 219], [134, 311]]
10000
100000
[[2227, 448], [1576, 1186]]
400
1
[[261, 126], [97, 295]]
[[215, 140], [92, 267]]
[[269, 134], [81, 313]]
[[415, 221], [67, 556]]
[[258, 158], [90, 380]]
[[309, 169], [178, 346]]
[[186, 120], [73, 229]]
[[1371, 926], [522, 1744]]
[[3098, 1127], [1874, 3901]]
5
...
4
0.6999
[[404, 199], [47, 523]]
0.5028
[[212, 156], [70, 279]]
[[228, 199], [156, 289]]
10000
100000
[[1727, 948], [605, 2157]]
100
1
[[231, 156], [69, 323]]
[[151, 204], [20, 339]]
[[239, 164], [67, 327]]
[[377, 259], [38, 585]]
[[222, 194], [45, 425]]
[[215, 263], [61, 463]]
[[111, 195], [5, 297]]
[[795, 1502], [72, 2194]]
[[2230, 372], [2742, 4656]]
5
...
4
0.6886
[[238, 365], [13, 557]]
0.5028
[[139, 229], [9, 340]]
[[121, 306], [21, 424]]
10000
100000
[[1435, 1240], [300, 2462]]
200
1
[[242, 145], [58, 334]]
[[167, 188], [29, 330]]
[[260, 143], [66, 328]]
[[391, 245], [51, 572]]
[[248, 168], [46, 424]]
[[243, 235], [84, 440]]
[[174, 132], [46, 256]]
[[1238, 1059], [286, 1980]]
[[2789, 620], [2183, 4408]]
5
...
4
0.7197
[[382, 221], [27, 543]]
0.5028
[[208, 160], [47, 302]]
[[174, 253], [64, 381]]
10000
100000
[[1551, 1124], [334, 2428]]
300
1
[[302, 85], [217, 175]]
[[211, 144], [101, 258]]
[[320, 83], [191, 203]]
[[500, 136], [150, 473]]
[[321, 95], [194, 276]]
[[320, 158], [232, 292]]
[[198, 108], [77, 225]]
[[1444, 853], [500, 1766]]
[[3418, 1585], [1554, 3443]]
5
...
4
0.6861
[[421, 182], [42, 528]]
0.5028
[[234, 134], [83, 266]]
[[232, 195], [141, 304]]
10000
100000
[[1974, 701], [1085, 1677]]
400
1
[[245, 142], [83, 309]]
[[171, 184], [47, 312]]
[[279, 124], [87, 307]]
[[410, 226], [57, 566]]
[[266, 150], [82, 388]]
[[282, 196], [135, 389]]
[[115, 191], [15, 287]]
[[879, 1418], [146, 2120]]
[[2532, 637], [2440, 4391]]
5
...
5
0.6923
[[254, 349], [14, 556]]
0.5028
[[158, 210], [15, 334]]
[[135, 292], [44, 401]]
10000
100000
[[1653, 1022], [491, 2271]]
100
1
[[276, 111], [129, 263]]
[[179, 176], [52, 307]]
[[297, 106], [114, 280]]
[[456, 180], [88, 535]]
[[292, 124], [131, 339]]
[[283, 195], [145, 379]]
[[114, 192], [9, 293]]
[[841, 1456], [81, 2185]]
[[2624, 740], [2348, 4288]]
5
...
5
0.6912
[[260, 343], [10, 560]]
0.5028
[[146, 222], [12, 337]]
[[115, 312], [23, 422]]
10000
100000
[[1783, 892], [659, 2103]]
200
1
[[199, 188], [41, 351]]
[[146, 209], [10, 349]]
[[209, 194], [43, 351]]
[[324, 312], [29, 594]]
[[196, 220], [24, 446]]
[[198, 280], [27, 497]]
[[165, 141], [40, 262]]
[[1149, 1148], [242, 2024]]
[[2421, 416], [2551, 4612]]
5
...
5
0.7033
[[341, 262], [30, 540]]
0.5028
[[197, 171], [37, 312]]
[[161, 266], [58, 387]]
10000
100000
[[1272, 1403], [174, 2588]]
300
1
[[373, 14], [351, 41]]
[[338, 17], [298, 61]]
[[389, 14], [335, 59]]
[[611, 25], [439, 184]]
[[398, 18], [408, 62]]
[[460, 18], [465, 59]]
[[269, 37], [212, 90]]
[[2030, 267], [1354, 912]]
[[4599, 3650], [373, 1378]]
5
...
5
0.5977
[[532, 71], [162, 408]]
0.5028
[[326, 42], [216, 133]]
[[381, 46], [356, 89]]
10000
100000
[[2569, 106], [2296, 466]]
400
1
[[245, 142], [72, 320]]
[[158, 197], [27, 332]]
[[267, 136], [79, 315]]
[[396, 240], [46, 577]]
[[247, 169], [68, 402]]
[[241, 237], [90, 434]]
[[100, 206], [4, 298]]
[[752, 1545], [49, 2217]]
[[2306, 431], [2666, 4597]]
5
...
6
0.6903
[[230, 373], [5, 565]]
0.5028
[[128, 240], [2, 347]]
[[101, 326], [18, 427]]
10000
100000
[[1554, 1121], [382, 2380]]
100
1
[[197, 190], [32, 360]]
[[138, 217], [9, 350]]
[[220, 183], [18, 376]]
[[324, 312], [29, 594]]
[[216, 200], [21, 449]]
[[214, 264], [39, 485]]
[[133, 173], [23, 279]]
[[961, 1336], [117, 2149]]
[[2270, 265], [2702, 4763]]
5
...
6
0.7033
[[315, 288], [11, 559]]
0.5028
[[159, 209], [15, 334]]
[[113, 314], [18, 427]]
10000
100000
[[1309, 1366], [148, 2614]]
200
1
[[162, 225], [13, 379]]
[[131, 224], [6, 353]]
[[165, 238], [5, 389]]
[[244, 392], [21, 602]]
[[165, 251], [14, 456]]
[[187, 291], [21, 503]]
[[175, 131], [43, 259]]
[[1190, 1107], [263, 2003]]
[[2244, 343], [2728, 4685]]
5
...
6
0.6929
[[368, 235], [27, 543]]
0.5028
[[194, 174], [41, 308]]
[[168, 259], [54, 391]]
10000
100000
[[1054, 1621], [80, 2682]]
300
1
[[255, 132], [128, 264]]
[[214, 141], [86, 273]]
[[280, 123], [114, 280]]
[[418, 218], [69, 554]]
[[267, 149], [114, 356]]
[[296, 182], [187, 337]]
[[248, 58], [207, 95]]
[[1799, 498], [1345, 921]]
[[3529, 2043], [1443, 2985]]
5
...
3
0.6514
[[514, 89], [325, 245]]
0.5028
[[298, 70], [196, 153]]
[[290, 137], [252, 193]]
10000
100000
[[1730, 945], [698, 2064]]
5
1
[[254, 133], [107, 285]]
[[201, 154], [73, 286]]
[[272, 131], [111, 283]]
[[414, 222], [71, 552]]
[[261, 155], [99, 371]]
[[283, 195], [146, 378]]
[[109, 197], [11, 291]]
[[800, 1497], [165, 2101]]
[[2485, 772], [2487, 4256]]
5
...
3
0.6741
[[212, 391], [11, 559]]
0.5028
[[141, 227], [21, 328]]
[[137, 290], [60, 385]]
10000
100000
[[1685, 990], [607, 2155]]
25
1
[[163, 224], [37, 355]]
[[174, 181], [71, 288]]
[[177, 226], [27, 367]]
[[266, 370], [33, 590]]
[[171, 245], [45, 425]]
[[225, 253], [91, 433]]
[[230, 76], [180, 122]]
[[1743, 554], [1049, 1217]]
[[2919, 1353], [2053, 3675]]
5
...
3
0.6594
[[512, 91], [128, 442]]
0.5028
[[283, 85], [170, 179]]
[[277, 150], [268, 177]]
10000
100000
[[1176, 1499], [304, 2458]]
50
1
[[224, 163], [88, 304]]
[[206, 149], [86, 273]]
[[249, 154], [64, 330]]
[[376, 260], [62, 561]]
[[246, 170], [86, 384]]
[[274, 204], [134, 390]]
[[200, 106], [100, 202]]
[[1504, 793], [635, 1631]]
[[3079, 1155], [1893, 3873]]
5
...
3
0.6952
[[446, 157], [51, 519]]
0.5028
[[241, 127], [93, 256]]
[[237, 190], [170, 275]]
10000
100000
[[1575, 1100], [520, 2242]]
75
1
[[313, 74], [253, 139]]
[[240, 115], [143, 216]]
[[336, 67], [211, 183]]
[[530, 106], [207, 416]]
[[322, 94], [218, 252]]
[[350, 128], [273, 251]]
[[150, 156], [46, 256]]
[[1131, 1166], [404, 1862]]
[[3222, 1709], [1750, 3319]]
5
...
4
0.6541
[[297, 306], [24, 546]]
0.5028
[[182, 186], [55, 294]]
[[211, 216], [126, 319]]
10000
100000
[[2091, 584], [1305, 1457]]
5
1
[[223, 164], [74, 318]]
[[186, 169], [72, 287]]
[[255, 148], [80, 314]]
[[375, 261], [52, 571]]
[[239, 177], [70, 400]]
[[263, 215], [104, 420]]
[[167, 139], [48, 254]]
[[1214, 1083], [388, 1878]]
[[2755, 840], [2217, 4188]]
5
...
4
0.6943
[[350, 253], [25, 545]]
0.5028
[[193, 175], [65, 284]]
[[201, 226], [121, 324]]
10000
100000
[[1541, 1134], [452, 2310]]
25
1
[[291, 96], [196, 196]]
[[204, 151], [97, 262]]
[[316, 87], [169, 225]]
[[495, 141], [137, 486]]
[[306, 110], [170, 300]]
[[306, 172], [194, 330]]
[[136, 170], [24, 278]]
[[958, 1339], [180, 2086]]
[[2876, 1143], [2096, 3885]]
5
...
4
0.6761
[[283, 320], [14, 556]]
0.5028
[[158, 210], [18, 331]]
[[147, 280], [54, 391]]
10000
100000
[[1918, 757], [963, 1799]]
50
1
[[167, 220], [24, 368]]
[[174, 181], [47, 312]]
[[204, 199], [22, 372]]
[[281, 355], [40, 583]]
[[188, 228], [33, 437]]
[[230, 248], [76, 448]]
[[221, 85], [156, 146]]
[[1617, 680], [853, 1413]]
[[2861, 1095], [2111, 3933]]
5
...
4
0.6794
[[474, 129], [114, 456]]
0.5028
[[273, 95], [131, 218]]
[[248, 179], [193, 252]]
10000
100000
[[1244, 1431], [242, 2520]]
75
1
[[282, 105], [156, 236]]
[[209, 146], [87, 272]]
[[293, 110], [123, 271]]
[[448, 188], [88, 535]]
[[278, 138], [120, 350]]
[[291, 187], [177, 347]]
[[180, 126], [84, 218]]
[[1328, 969], [507, 1759]]
[[3129, 1258], [1843, 3770]]
5
...
5
0.6899
[[382, 221], [38, 532]]
0.5028
[[222, 146], [67, 282]]
[[207, 220], [146, 299]]
10000
100000
[[1801, 874], [751, 2011]]
5
1
[[279, 108], [173, 219]]
[[202, 153], [78, 281]]
[[307, 96], [154, 240]]
[[464, 172], [116, 507]]
[[291, 125], [139, 331]]
[[297, 181], [168, 356]]
[[214, 92], [133, 169]]
[[1536, 761], [745, 1521]]
[[3376, 1573], [1596, 3455]]
5
...
5
0.6831
[[465, 138], [89, 481]]
0.5028
[[250, 118], [114, 235]]
[[226, 201], [173, 272]]
10000
100000
[[1840, 835], [828, 1934]]
25
1
[[180, 207], [27, 365]]
[[135, 220], [29, 330]]
[[201, 202], [22, 372]]
[[274, 362], [28, 595]]
[[181, 235], [19, 451]]
[[197, 281], [46, 478]]
[[149, 157], [26, 276]]
[[1071, 1226], [190, 2076]]
[[2239, 361], [2733, 4667]]
5
...
5
0.6906
[[343, 260], [20, 550]]
0.5028
[[171, 197], [25, 324]]
[[149, 278], [45, 400]]
10000
100000
[[1168, 1507], [171, 2591]]
50
1
[[142, 245], [15, 377]]
[[131, 224], [23, 336]]
[[150, 253], [12, 382]]
[[229, 407], [25, 598]]
[[162, 254], [18, 452]]
[[187, 291], [42, 482]]
[[230, 76], [153, 149]]
[[1607, 690], [953, 1313]]
[[2608, 1088], [2364, 3940]]
5
...
5
0.6548
[[482, 121], [216, 354]]
0.5028
[[264, 104], [142, 207]]
[[237, 190], [171, 274]]
10000
100000
[[1001, 1674], [135, 2627]]
75
1
[[305, 82], [211, 181]]
[[223, 132], [121, 238]]
[[326, 77], [175, 219]]
[[501, 135], [123, 500]]
[[327, 89], [186, 284]]
[[337, 141], [241, 283]]
[[181, 125], [69, 233]]
[[1321, 976], [473, 1793]]
[[3340, 1530], [1632, 3498]]
5
...
6
0.6838
[[380, 223], [35, 535]]
0.5028
[[214, 154], [79, 270]]
[[218, 209], [140, 305]]
10000
100000
[[2019, 656], [1057, 1705]]
5
1
[[230, 157], [78, 314]]
[[195, 160], [62, 297]]
[[256, 147], [87, 307]]
[[388, 248], [47, 576]]
[[241, 175], [72, 398]]
[[274, 204], [122, 402]]
[[131, 175], [30, 272]]
[[986, 1311], [252, 2014]]
[[2570, 720], [2402, 4308]]
5
...
6
0.6878
[[272, 331], [22, 548]]
0.5028
[[173, 195], [38, 311]]
[[162, 265], [80, 365]]
10000
100000
[[1584, 1091], [468, 2294]]
25
1
[[266, 121], [118, 274]]
[[188, 167], [62, 297]]
[[285, 118], [100, 294]]
[[426, 210], [66, 557]]
[[279, 137], [98, 372]]
[[283, 195], [157, 367]]
[[182, 124], [77, 225]]
[[1293, 1004], [418, 1848]]
[[3020, 1019], [1952, 4009]]
5
...
6
0.7029
[[398, 205], [38, 532]]
0.5028
[[209, 159], [53, 296]]
[[175, 252], [95, 350]]
10000
100000
[[1727, 948], [601, 2161]]
50
45 rows × 23 columns
In [64]:
x = df['width']
y = df['num_layers']
z = df['overall_acc']
# TODO: 3D PLOT
print(x[:5])
print(y[:5])
print(z[:5])
1 5
1 25
1 50
1 75
1 5
Name: width, dtype: int64
1 3
1 3
1 3
1 3
1 4
Name: num_layers, dtype: int64
1 0.6514
1 0.6741
1 0.6594
1 0.6952
1 0.6541
Name: overall_acc, dtype: float64
In [65]:
x = df['width']
y = df['overall_acc']
plt.scatter(x,y)
plt.xlabel('width')
plt.ylabel('accuracy')
plt.ylim((0.0,1.0))
Out[65]:
(0.0, 1.0)
In [66]:
x = df['num_layers']
y = df['overall_acc']
plt.scatter(x,y)
plt.xlabel('num_layers')
plt.ylabel('accuracy')
plt.ylim((0.0,1.0))
Out[66]:
(0.0, 1.0)
In [ ]:
In [ ]:
In [91]:
X = df['num_layers'].unique()
Y = df['width'].unique()
X, Y = np.meshgrid(X, Y)
print(Y.shape)
print(X.shape)
(8, 4)
(8, 4)
In [94]:
Z = np.zeros(X.shape)
for i in range(X.shape[0]):
new_row = []
for j in range(X.shape[1]):
#get overall accuracy for the width and depth
#print((x, y))
print((i,j))
target = df[(df['num_layers'] == X[i][j]) & (df['width'] == Y[i][j])]
if (len(target) > 0):
Z[i][j] = target.head(1).overall_acc[1]
Z
(0, 0)
(0, 1)
(0, 2)
(0, 3)
(1, 0)
(1, 1)
(1, 2)
(1, 3)
(2, 0)
(2, 1)
(2, 2)
(2, 3)
(3, 0)
(3, 1)
(3, 2)
(3, 3)
(4, 0)
(4, 1)
(4, 2)
(4, 3)
(5, 0)
(5, 1)
(5, 2)
(5, 3)
(6, 0)
(6, 1)
(6, 2)
(6, 3)
(7, 0)
(7, 1)
(7, 2)
(7, 3)
Out[94]:
array([[ 0.6514, 0.6541, 0.6899, 0.6838],
[ 0.6741, 0.6943, 0.6831, 0.6878],
[ 0.6594, 0.6761, 0.6906, 0.7029],
[ 0.6952, 0.6794, 0.6548, 0. ],
[ 0.6851, 0.6999, 0.6923, 0.6903],
[ 0.6293, 0.6886, 0.6912, 0.7033],
[ 0.6786, 0.7197, 0.7033, 0.6929],
[ 0.6516, 0.6861, 0.5977, 0. ]])
In [95]:
Z[3][3] = .7
Z[7][3] = .7
Z
Out[95]:
array([[ 0.6514, 0.6541, 0.6899, 0.6838],
[ 0.6741, 0.6943, 0.6831, 0.6878],
[ 0.6594, 0.6761, 0.6906, 0.7029],
[ 0.6952, 0.6794, 0.6548, 0.7 ],
[ 0.6851, 0.6999, 0.6923, 0.6903],
[ 0.6293, 0.6886, 0.6912, 0.7033],
[ 0.6786, 0.7197, 0.7033, 0.6929],
[ 0.6516, 0.6861, 0.5977, 0.7 ]])
In [110]:
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
import matplotlib.pyplot as plt
import numpy as np
fig = plt.figure(figsize=(16,12))
ax = fig.gca(projection='3d')
surf = ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cm.coolwarm,
linewidth=0, antialiased=False)
ax.set_xlabel('Depth', fontsize=20)
ax.set_ylabel('Width', fontsize=20)
ax.set_zlabel("Accuracy", fontsize=20)
plt.setp(ax.get_xticklabels(), fontsize=15)
plt.setp(ax.get_yticklabels(), fontsize=15)
plt.setp(ax.get_zticklabels(), fontsize=15)
ax.set_title("Width and Depth of Network vs. Accuracy", fontsize=30)
Out[110]:
<matplotlib.text.Text at 0x7f6ae4187290>
In [81]:
np.zeros(X.shape)
Out[81]:
array([[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.],
[ 0., 0., 0., 0.]])
In [ ]:
Content source: Kkevsterrr/cicile-engine
Similar notebooks: