dima806/popular_sneakers_detection

Detects popular sneaker type with 69% accuracy based on an image.

See this Kaggle notebook for details.

Classification report:

                                     precision    recall  f1-score   support

                  adidas_forum_high     0.7097    0.7333    0.7213        60
                   adidas_forum_low     0.8056    0.4833    0.6042        60
                     adidas_gazelle     0.6098    0.4167    0.4950        60
                      adidas_nmd_r1     0.8596    0.8167    0.8376        60
                       adidas_samba     0.5225    0.9667    0.6784        60
                  adidas_stan_smith     0.5974    0.7667    0.6715        60
                   adidas_superstar     0.7561    0.5167    0.6139        60
                  adidas_ultraboost     0.7692    0.8333    0.8000        60
                 asics_gel-lyte_iii     0.8308    0.9000    0.8640        60
             converse_chuck_70_high     0.7458    0.7333    0.7395        60
              converse_chuck_70_low     0.6250    0.5000    0.5556        60
converse_chuck_taylor_all-star_high     0.7429    0.8667    0.8000        60
 converse_chuck_taylor_all-star_low     0.5714    0.8000    0.6667        60
                  converse_one_star     0.9038    0.7833    0.8393        60
                    new_balance_327     0.8824    0.7500    0.8108        60
                    new_balance_550     0.7400    0.6167    0.6727        60
                    new_balance_574     0.8780    0.6000    0.7129        60
                    new_balance_990     0.5952    0.8333    0.6944        60
                    new_balance_992     0.9333    0.2333    0.3733        60
              nike_air_force_1_high     0.3654    0.6333    0.4634        60
               nike_air_force_1_low     0.4500    0.3000    0.3600        60
               nike_air_force_1_mid     0.4545    0.1667    0.2439        60
                 nike_air_jordan_11     0.7222    0.8667    0.7879        60
             nike_air_jordan_1_high     0.5657    0.9333    0.7044        60
              nike_air_jordan_1_low     0.8125    0.4333    0.5652        60
                  nike_air_jordan_3     0.5904    0.8167    0.6853        60
                  nike_air_jordan_4     0.8000    0.6000    0.6857        60
                     nike_air_max_1     0.5652    0.8667    0.6842        60
                   nike_air_max_270     0.8980    0.7333    0.8073        60
                    nike_air_max_90     0.8667    0.4333    0.5778        60
                    nike_air_max_95     0.7500    0.7000    0.7241        60
                    nike_air_max_97     0.7895    1.0000    0.8824        60
             nike_air_max_plus_(tn)     0.7451    0.6333    0.6847        60
          nike_air_vapormax_flyknit     0.8491    0.7500    0.7965        60
             nike_air_vapormax_plus     0.7458    0.7333    0.7395        60
                 nike_blazer_mid_77     0.7797    0.7667    0.7731        60
                        nike_cortez     0.5253    0.8667    0.6541        60
                     nike_dunk_high     0.5128    0.3333    0.4040        60
                      nike_dunk_low     0.6164    0.7500    0.6767        60
                 puma_suede_classic     0.7302    0.7667    0.7480        60
             reebok_classic_leather     0.8261    0.3167    0.4578        60
                   reebok_club_c_85     0.4865    0.6000    0.5373        60
                       salomon_xt-6     1.0000    0.7333    0.8462        60
                     vans_authentic     0.8462    0.7333    0.7857        60
                     vans_old_skool     0.5663    0.7833    0.6573        60
                        vans_sk8-hi     0.7750    0.5167    0.6200        60
          vans_slip-on_checkerboard     0.8507    0.9500    0.8976        60
              yeezy_700_wave_runner     0.7273    0.9333    0.8175        60
                 yeezy_boost_350_v2     0.9333    0.9333    0.9333        60
                        yeezy_slide     0.7867    0.9833    0.8741        60

                           accuracy                         0.6943      3000
                          macro avg     0.7202    0.6943    0.6845      3000
                       weighted avg     0.7202    0.6943    0.6845      3000