Philippe Tarral: raw computed results
The raw ranking from the production similarity calculation (face-embedding centroid cosine → gender-cohort z-normalization). The web-opinion crown, curation, and fusion re-rank are removed.
- 1
81.6%z 4.15 - 2
63.1%z 3.83 - 3
59.5%z 3.79 - 4
50.7%z 3.70 - 5
49.1%z 3.68 - 6
47.9%z 3.66 - 7
45.7%z 3.52 - 8
44.8%z 3.46 - 9
44.3%z 3.43 - 10
44%z 3.41 - 11
43.4%z 3.37 - 12
43.1%z 3.35 - 13
41.3%z 3.23 - 14
41%z 3.21 - 15
41%z 3.21 - 16
41%z 3.21 - 17
40.5%z 3.18 - 18
40.4%z 3.17 - 19
40.1%z 3.16 - 20
40.1%z 3.16 - 21
39.4%z 3.11 - 22
39.2%z 3.10 - 23
39%z 3.08 - 24
38.9%z 3.08 - 25
38.2%z 3.03 - 26
37.4%z 2.98 - 27
37.4%z 2.98 - 28
37.4%z 2.98 - 29
37.4%z 2.98 - 30
36.9%z 2.95 - 31
36.9%z 2.95 - 32
36.9%z 2.95 - 33
36.8%z 2.94 - 34
36.7%z 2.94 - 35
36.6%z 2.93 - 36
36.3%z 2.91 - 37
36.1%z 2.89 - 38
35.9%z 2.88 - 39
35.7%z 2.87 - 40
35.5%z 2.86 - 41
35.5%z 2.86 - 42
35.5%z 2.85 - 43
35.4%z 2.85 - 44
35.4%z 2.85 - 45
35.3%z 2.84 - 46
35.3%z 2.84 - 47
34.9%z 2.82 - 48
34.9%z 2.82 - 49
34.9%z 2.82 - 50
34.7%z 2.80