2025-06-29 21:11:59 | WARNING | dataFuture:get_latest_valid_dir:99 - 使用的数据目录日期(2025-06-27)不是今天(2025-06-29)，可能不是最新数据
2025-06-29 21:12:01 | INFO | model_gen:main:87 - 性能监控状态: False
2025-06-29 21:12:01 | INFO | model_gen:main:90 - {'config': {'envs': {'env': {'mode': 'prod', 'enable_monitor': False, 'n_jobs': 1}, 'incremental': {'enabled': True, 'lookback_days': 20, 'start_date': None, 'end_date': 'today'}, 'run': {'n_jobs': 1, 'retry_times': 3, 'retry_delay': 2}, 'shard': {'num_shards': 1, 'shard_id': 0}}, 'strategy': {'time_bar': 5, 'pred_periods': 12, 'base_key': 'B8Wstats250412', 'task_config': {'selected_expr_mapping': 'B8Wstats250412', 'config_key': 'B8Wstats250412', 'combo_key': 'combo52', 'metric_key': ['B8Wstats250412'], 'versions_to_optimize': ['v7', 'v8']}}, 'compression': {'type': 'zstd', 'level': 3}, 'files': {'base_dir': 'output', 'data': 'data_5min.parquet', 'pctchg': 'pctchg_5min_pred12.parquet'}, 'feature': {'include_time_comd': True, 'additional_columns': []}, 'epoch': {'freq': 'W'}}}
2025-06-29 21:12:02 | INFO | model_gen:main:95 - {'action': 'params_initialized', 'config_key': 'B8Wstats250412', 'combo_key': 'combo52', 'time_bar': 5, 'pred_periods': 12, 'pctchg_path': 'pctchg_5min_pred12.parquet'}
2025-06-29 21:12:02 | INFO | model_gen:main:99 - {'action': 'epoch_list_generated', 'count': 3}
2025-06-29 21:12:02 | INFO | model_gen:main:108 - {'action': 'shard_tasks', 'count': 3, 'head': array([[0, Period('2025-06-09/2025-06-15', 'W-SUN')],
       [0, Period('2025-06-16/2025-06-22', 'W-SUN')],
       [0, Period('2025-06-23/2025-06-29', 'W-SUN')]], dtype=object)}
2025-06-29 21:12:02 | INFO | model_gen:main:110 - {'action': 'parallel_start', 'n_jobs': 1, 'retry_times': 3}
2025-06-29 22:59:03 | INFO | model_gen:main:127 - {'action': 'parallel_done', 'success_count': 3, 'fail_count': 0}
2025-06-29 22:59:03 | INFO | model_gen:main:132 - {'action': 'success', 'description': 'model_gen 完成'}
sort done
内存优化: 12566.18 MB -> 6496.35 MB, 节省 48.30%
训练数据处理耗时: 243.2274秒
sort done
内存优化: 12566.18 MB -> 6496.35 MB, 节省 48.30%
预测数据处理耗时: 1388.0256秒
使用顺序模式运行 1 个模型...
应用组合函数: LGBMRegressor
[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 32.366670 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 18853
[LightGBM] [Info] Number of data points in the train set: 8637281, number of used features: 191
[LightGBM] [Info] Start training from score 0.001055
train_and_evaluate_pl: best_score {'TA': 0.009305029610683515, 'TH': 0.00679939427515843, 'TQ': 0.008446304998927392, 'autocorr': 0.9268665315761205}
模型训练和预测耗时: 553.8489秒
结果最后日期: 2025-06-21 01:00:00
结果保存耗时: 0.0175秒
every thing is done
sort done
内存优化: 12594.36 MB -> 6510.92 MB, 节省 48.30%
训练数据处理耗时: 337.8152秒
sort done
内存优化: 12594.36 MB -> 6510.92 MB, 节省 48.30%
预测数据处理耗时: 983.2782秒
使用顺序模式运行 1 个模型...
应用组合函数: LGBMRegressor
[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 27.972648 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 18860
[LightGBM] [Info] Number of data points in the train set: 8658411, number of used features: 191
[LightGBM] [Info] Start training from score 0.001120
train_and_evaluate_pl: best_score {'TA': 0.009187992449692718, 'TH': 0.006707988650856393, 'TQ': 0.008333848121089945, 'autocorr': 0.9253819051617937}
模型训练和预测耗时: 596.4582秒
结果最后日期: 2025-06-27 15:15:00
结果保存耗时: 0.1650秒
every thing is done
sort done
内存优化: 12594.36 MB -> 6510.92 MB, 节省 48.30%
训练数据处理耗时: 266.6399秒
sort done
内存优化: 12594.36 MB -> 6510.92 MB, 节省 48.30%
预测数据处理耗时: 1346.3681秒
使用顺序模式运行 1 个模型...
应用组合函数: LGBMRegressor
[LightGBM] [Info] Auto-choosing col-wise multi-threading, the overhead of testing was 41.276749 seconds.
You can set `force_col_wise=true` to remove the overhead.
[LightGBM] [Info] Total Bins 18860
[LightGBM] [Info] Number of data points in the train set: 8678567, number of used features: 191
[LightGBM] [Info] Start training from score 0.001120
train_and_evaluate_pl: best_score {'TA': 0.00915262816767101, 'TH': 0.006671060896365465, 'TQ': 0.008299231053183866, 'autocorr': 0.925890213448328}
模型训练和预测耗时: 603.5877秒
结果最后日期: 2025-06-27 15:15:00
结果保存耗时: 0.0383秒
every thing is done