-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathflightrank-version8.log
More file actions
206 lines (206 loc) · 12.9 KB
/
Copy pathflightrank-version8.log
File metadata and controls
206 lines (206 loc) · 12.9 KB
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
70.9s 1 /usr/local/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
70.9s 2 from .autonotebook import tqdm as notebook_tqdm
70.9s 3
70.9s 4 /usr/local/lib/python3.10/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html
70.9s 5 from .autonotebook import tqdm as notebook_tqdm
139.1s 6 Adding advanced business traveler features...
144.0s 7 Advanced business traveler features added successfully!
144.0s 8 Key improvements:
144.0s 9 - Corporate policy compliance and tax efficiency
144.0s 10 - Advanced time preferences for business travelers
144.0s 11 - Seat scarcity and availability features
144.0s 12 - Cancellation/exchange policy flexibility
144.0s 13 - Premium carrier and aircraft preferences
144.0s 14 - Hub airport and route optimization
144.0s 15 - Strategic business value interactions
144.0s 16 - Safe feature creation with proper dependency ordering
144.0s 17 - Targeting 0.5+ Kaggle score with business insights
148.2s 18 Using 180 features (35 categorical)
148.2s 19 CatBoost categorical indices: 35 features
148.2s 20 Performing FIXED Early CatBoost Validation...
148.2s 21 catboost_cat_indices found: 35 categorical features
148.2s 22 Indices: [0, 1, 2, 3, 7, 8, 9, 13, 15, 16]...
148.2s 23 All 35 categorical features found in dataset
148.2s 24 Testing CatBoost Pool creation with EXACT SAME logic as actual training...
148.2s 25 Applying EXACT SAME unified categorical encoding as actual training...
148.2s 26 Processing categorical feature: nationality
148.3s 27 Processing categorical feature: searchRoute
148.3s 28 Processing categorical feature: corporateTariffCode
148.3s 29 Processing categorical feature: bySelf
148.3s 30 Processing categorical feature: sex
148.3s 31 Processing categorical feature: companyID
148.3s 32 Processing categorical feature: legs0_segments0_aircraft_code
148.3s 33 Processing categorical feature: legs0_segments0_arrivalTo_airport_city_iata
148.3s 34 Processing categorical feature: legs0_segments0_arrivalTo_airport_iata
148.3s 35 Processing categorical feature: legs0_segments0_departureFrom_airport_iata
148.3s 36 Processing categorical feature: legs0_segments0_marketingCarrier_code
148.3s 37 Processing categorical feature: legs0_segments0_operatingCarrier_code
148.3s 38 Processing categorical feature: legs0_segments0_flightNumber
148.3s 39 Processing categorical feature: legs0_segments1_aircraft_code
148.3s 40 Processing categorical feature: legs0_segments1_arrivalTo_airport_city_iata
148.3s 41 Processing categorical feature: legs0_segments1_arrivalTo_airport_iata
148.3s 42 Processing categorical feature: legs0_segments1_departureFrom_airport_iata
148.3s 43 Processing categorical feature: legs0_segments1_marketingCarrier_code
148.3s 44 Processing categorical feature: legs0_segments1_operatingCarrier_code
148.4s 45 Processing categorical feature: legs0_segments1_flightNumber
148.4s 46 Processing categorical feature: legs1_segments0_aircraft_code
148.4s 47 Processing categorical feature: legs1_segments0_arrivalTo_airport_city_iata
148.4s 48 Processing categorical feature: legs1_segments0_arrivalTo_airport_iata
148.4s 49 Processing categorical feature: legs1_segments0_departureFrom_airport_iata
148.4s 50 Processing categorical feature: legs1_segments0_marketingCarrier_code
148.4s 51 Processing categorical feature: legs1_segments0_operatingCarrier_code
148.4s 52 Processing categorical feature: legs1_segments0_flightNumber
148.4s 53 Processing categorical feature: legs1_segments1_aircraft_code
148.4s 54 Processing categorical feature: legs1_segments1_arrivalTo_airport_city_iata
148.4s 55 Processing categorical feature: legs1_segments1_arrivalTo_airport_iata
148.4s 56 Processing categorical feature: legs1_segments1_departureFrom_airport_iata
148.4s 57 Processing categorical feature: legs1_segments1_marketingCarrier_code
148.4s 58 Processing categorical feature: legs1_segments1_operatingCarrier_code
148.4s 59 Processing categorical feature: legs1_segments1_flightNumber
148.4s 60 Processing categorical feature: price_bucket
149.1s 61 Sample data shape: (1000, 180)
149.1s 62 Categorical indices: [0, 1, 2, 3, 7]...
149.1s 63 Cat feature 0 (col 0, 'bySelf'): dtype=int32, sample=[0, 0, 0]
149.1s 64 Cat feature 1 (col 1, 'companyID'): dtype=int32, sample=[7, 7, 7]
149.1s 65 Cat feature 2 (col 2, 'corporateTariffCode'): dtype=int32, sample=[0, 7, 0]
149.1s 66 CatBoost Pool created successfully with EXACT SAME logic as training!
149.1s 67 Sample size: 1000 rows
149.1s 68 Features: 180 columns
149.1s 69 Categorical features: 35 indices
149.1s 70 Data types verified: mixed int32/float32 compatible with CatBoost
150.0s 71 CatBoost model training test successful with EXACT SAME logic!
150.0s 72 Predictions shape: (1000,)
150.0s 73 Sample predictions: [ 0.85018012 0.17829717 -0.08035397 -0.0805669 -0.0805669 ]
150.0s 74 No data type conflicts detected!
150.0s 75
150.0s 76 Early CatBoost Validation PASSED with FIXED mixed-type handling!
150.0s 77 All CatBoost components verified and working
150.0s 78 Data type processing EXACTLY matches actual training
150.0s 79 No late-stage errors expected during actual training
150.0s 80 Safe to proceed with full model training pipeline
150.0s 81 ============================================================
214.3s 82
214.3s 83 Training XGBoost with v2 business-optimized parameters...
214.3s 84 V2 optimizations for better convergence:
214.3s 85 - Reduced depth (8) to prevent slight overfitting
214.3s 86 - Improved sampling rates for better generalization
214.3s 87 - Fine-tuned regularization for optimal expressiveness
214.3s 88 - Adjusted learning rate for faster convergence
228.0s 89 [0] train-ndcg@3:0.78564 val-ndcg@3:0.80987
285.3s 90 [50] train-ndcg@3:0.81576 val-ndcg@3:0.83376
346.2s 91 [100] train-ndcg@3:0.82366 val-ndcg@3:0.83657
410.1s 92 [150] train-ndcg@3:0.83067 val-ndcg@3:0.83838
474.5s 93 [200] train-ndcg@3:0.83721 val-ndcg@3:0.83973
538.6s 94 [250] train-ndcg@3:0.84342 val-ndcg@3:0.84100
602.0s 95 [300] train-ndcg@3:0.84874 val-ndcg@3:0.84202
660.7s 96 [350] train-ndcg@3:0.85386 val-ndcg@3:0.84297
720.9s 97 [400] train-ndcg@3:0.85807 val-ndcg@3:0.84337
776.6s 98 [450] train-ndcg@3:0.86272 val-ndcg@3:0.84341
831.7s 99 [500] train-ndcg@3:0.86614 val-ndcg@3:0.84329
891.6s 100 [550] train-ndcg@3:0.86971 val-ndcg@3:0.84374
941.9s 101 [600] train-ndcg@3:0.87300 val-ndcg@3:0.84386
1002.2s 102 [650] train-ndcg@3:0.87562 val-ndcg@3:0.84408
1046.3s 103 [700] train-ndcg@3:0.87778 val-ndcg@3:0.84422
1101.6s 104 [750] train-ndcg@3:0.88063 val-ndcg@3:0.84426
1157.0s 105 [800] train-ndcg@3:0.88340 val-ndcg@3:0.84430
1206.1s 106 [850] train-ndcg@3:0.88546 val-ndcg@3:0.84451
1257.9s 107 [900] train-ndcg@3:0.88773 val-ndcg@3:0.84500
1312.2s 108 [950] train-ndcg@3:0.89052 val-ndcg@3:0.84522
1366.8s 109 [1000] train-ndcg@3:0.89354 val-ndcg@3:0.84484
1412.3s 110 [1042] train-ndcg@3:0.89669 val-ndcg@3:0.84534
1412.5s 111 Creating LightGBM Datasets...
1412.5s 112 LightGBM Datasets created successfully.
1414.4s 113
1414.4s 114 Training LightGBM model with v2 business-optimized parameters...
1414.4s 115 V2 business feature optimizations:
1414.4s 116 - Reduced leaves (140) and depth (10) to prevent overfitting
1414.4s 117 - Lower learning rate (0.15) for better stability
1414.4s 118 - Increased regularization (L1=0.006, L2=7.5) for generalization
1414.4s 119 - Conservative sampling for better model robustness
1541.7s 120 Training until validation scores don't improve for 130 rounds
1586.5s 121 [50] training's ndcg@3: 0.838298 valid_1's ndcg@3: 0.839591
1633.8s 122 [100] training's ndcg@3: 0.858681 valid_1's ndcg@3: 0.841336
1678.8s 123 [150] training's ndcg@3: 0.873883 valid_1's ndcg@3: 0.841788
1717.0s 124 [200] training's ndcg@3: 0.88456 valid_1's ndcg@3: 0.842278
1763.6s 125 [250] training's ndcg@3: 0.894915 valid_1's ndcg@3: 0.841154
1799.9s 126 [300] training's ndcg@3: 0.903706 valid_1's ndcg@3: 0.842088
1841.1s 127 [350] training's ndcg@3: 0.911544 valid_1's ndcg@3: 0.840891
1878.9s 128 [400] training's ndcg@3: 0.918839 valid_1's ndcg@3: 0.840583
1887.0s 129 Early stopping, best iteration is:
1887.0s 130 [282] training's ndcg@3: 0.900754 valid_1's ndcg@3: 0.842607
1887.8s 131 [LightGBM] [Fatal] Reducing `min_data_in_leaf` with `feature_pre_filter=true` may cause unexpected behaviour for features that were pre-filtered by the larger `min_data_in_leaf`.
1887.8s 132 You need to set `feature_pre_filter=false` to dynamically change the `min_data_in_leaf`.
1887.8s 133
1887.8s 134 --- Training LightGBM DART Model ---
1887.8s 135
1887.8s 136 Training Enhanced LightGBM DART model (v2)...
1887.8s 137 V2 DART optimizations:
1887.8s 138 - Refined learning rate (0.045) for better convergence
1887.8s 139 - Increased capacity (60 leaves) with depth control (12)
1887.8s 140 - Optimized DART dropout balance (0.075 rate, 0.48 skip)
1887.8s 141 - Added regularization for better generalization
1887.8s 142 [LightGBM] [Fatal] Reducing `min_data_in_leaf` with `feature_pre_filter=true` may cause unexpected behaviour for features that were pre-filtered by the larger `min_data_in_leaf`.
1887.8s 143 You need to set `feature_pre_filter=false` to dynamically change the `min_data_in_leaf`.
1887.8s 144
1887.8s 145 [LightGBM] [Fatal] Reducing `min_data_in_leaf` with `feature_pre_filter=true` may cause unexpected behaviour for features that were pre-filtered by the larger `min_data_in_leaf`.
1887.8s 146 You need to set `feature_pre_filter=false` to dynamically change the `min_data_in_leaf`.
1887.9s 147 [LightGBM] [Fatal] Reducing `min_data_in_leaf` with `feature_pre_filter=true` may cause unexpected behaviour for features that were pre-filtered by the larger `min_data_in_leaf`.
1887.9s 148 You need to set `feature_pre_filter=false` to dynamically change the `min_data_in_leaf`.
1889.1s 149 [LightGBM] [Fatal] Reducing `min_data_in_leaf` with `feature_pre_filter=true` may cause unexpected behaviour for features that were pre-filtered by the larger `min_data_in_leaf`.
1889.1s 150 You need to set `feature_pre_filter=false` to dynamically change the `min_data_in_leaf`.
1889.1s 151
1889.1s 152 [LightGBM] [Fatal] Reducing `min_data_in_leaf` with `feature_pre_filter=true` may cause unexpected behaviour for features that were pre-filtered by the larger `min_data_in_leaf`.
1889.1s 153 You need to set `feature_pre_filter=false` to dynamically change the `min_data_in_leaf`.
2021.8s 154 /usr/local/lib/python3.10/site-packages/lightgbm/callback.py:333: UserWarning: Early stopping is not available in dart mode
2021.8s 155 _log_warning("Early stopping is not available in dart mode")
2021.8s 156
2021.8s 157 /usr/local/lib/python3.10/site-packages/lightgbm/callback.py:333: UserWarning: Early stopping is not available in dart mode
2021.8s 158 _log_warning("Early stopping is not available in dart mode")
2056.8s 159 [50] valid_0's ndcg@3: 0.826599
2103.4s 160 [100] valid_0's ndcg@3: 0.829285
2154.4s 161 [150] valid_0's ndcg@3: 0.830514
2214.6s 162 [200] valid_0's ndcg@3: 0.831999
2270.8s 163 [250] valid_0's ndcg@3: 0.834289
2347.8s 164 [300] valid_0's ndcg@3: 0.834768
2425.1s 165 [350] valid_0's ndcg@3: 0.835424
2507.9s 166 [400] valid_0's ndcg@3: 0.836063
2595.0s 167 [450] valid_0's ndcg@3: 0.837309
2694.6s 168 [500] valid_0's ndcg@3: 0.836952
2817.8s 169 [550] valid_0's ndcg@3: 0.837339
2911.8s 170 [600] valid_0's ndcg@3: 0.838046
3024.5s 171 [650] valid_0's ndcg@3: 0.838453
3129.7s 172 [700] valid_0's ndcg@3: 0.839325
3235.4s 173 [750] valid_0's ndcg@3: 0.839923
3358.1s 174 [800] valid_0's ndcg@3: 0.840273
3468.6s 175 [850] valid_0's ndcg@3: 0.840311
3598.2s 176 [900] valid_0's ndcg@3: 0.840903
3713.8s 177 [950] valid_0's ndcg@3: 0.841412
3833.6s 178 [1000] valid_0's ndcg@3: 0.841571
3957.5s 179 [1050] valid_0's ndcg@3: 0.841679
4063.7s 180 [1100] valid_0's ndcg@3: 0.842276
4188.1s 181 [1150] valid_0's ndcg@3: 0.842249
4313.7s 182 [1200] valid_0's ndcg@3: 0.842248
4431.1s 183 [1250] valid_0's ndcg@3: 0.842623
4554.2s 184 [1300] valid_0's ndcg@3: 0.842513
4663.6s 185 [1350] valid_0's ndcg@3: 0.842651
4791.2s 186 [1400] valid_0's ndcg@3: 0.842489
4901.5s 187 [1450] valid_0's ndcg@3: 0.842229
5006.0s 188 [1500] valid_0's ndcg@3: 0.842906
5118.2s 189 [1550] valid_0's ndcg@3: 0.843373
5237.4s 190 [1600] valid_0's ndcg@3: 0.842931
5365.8s 191 [1650] valid_0's ndcg@3: 0.843108
5462.6s 192 [1700] valid_0's ndcg@3: 0.843577
5572.7s 193 [1750] valid_0's ndcg@3: 0.844182
5685.6s 194 [1800] valid_0's ndcg@3: 0.844068
5774.9s 195 [1850] valid_0's ndcg@3: 0.843944
5894.0s 196 [1900] valid_0's ndcg@3: 0.843679
6001.9s 197 [1950] valid_0's ndcg@3: 0.843263
6105.4s 198 [2000] valid_0's ndcg@3: 0.843664
6227.2s 199 [2050] valid_0's ndcg@3: 0.843839
6347.9s 200 [2100] valid_0's ndcg@3: 0.844111
6457.6s 201 [2150] valid_0's ndcg@3: 0.844119
6574.9s 202 [2200] valid_0's ndcg@3: 0.844405
6724.6s 203 [2250] valid_0's ndcg@3: 0.844383
6826.6s 204 [2300] valid_0's ndcg@3: 0.844701
6906.9s 205 [2350] valid_0's ndcg@3: 0.844149
7156.7s 206 [2400] valid_0's ndcg@3: 0.844231