ML之catboost：基于自定義數(shù)據(jù)集利用catboost 算法實現(xiàn)回歸預(yù)測(訓(xùn)練采用CPU和GPU兩種方式)

目錄

基于自定義數(shù)據(jù)集利用catboost 算法實現(xiàn)回歸預(yù)測(訓(xùn)練采用CPU和GPU兩種方式)

輸出結(jié)果

# T1、訓(xùn)練采用CPU

# T2、訓(xùn)練采用GPU

實現(xiàn)代碼

---

基于自定義數(shù)據(jù)集利用catboost 算法實現(xiàn)回歸預(yù)測(訓(xùn)練采用CPU和GPU兩種方式)

輸出結(jié)果

# T1、訓(xùn)練采用CPU

0: learn: 7.9608417 total: 50.2ms remaining: 1.46s 1: learn: 7.7618206 total: 50.5ms remaining: 707ms 2: learn: 7.5879985 total: 50.9ms remaining: 458ms 3: learn: 7.4148674 total: 51.4ms remaining: 334ms 4: learn: 7.2488388 total: 51.6ms remaining: 258ms 5: learn: 7.0676178 total: 51.8ms remaining: 207ms 6: learn: 6.8909273 total: 51.9ms remaining: 171ms 7: learn: 6.7186541 total: 52.1ms remaining: 143ms 8: learn: 6.5506878 total: 52.3ms remaining: 122ms 9: learn: 6.3869206 total: 52.4ms remaining: 105ms 10: learn: 6.2272476 total: 52.6ms remaining: 90.9ms 11: learn: 6.0715664 total: 52.8ms remaining: 79.2ms 12: learn: 5.9197772 total: 52.9ms remaining: 69.2ms 13: learn: 5.7717828 total: 53.1ms remaining: 60.7ms 14: learn: 5.6274882 total: 53.2ms remaining: 53.2ms 15: learn: 5.4868010 total: 53.4ms remaining: 46.7ms 16: learn: 5.3496310 total: 53.6ms remaining: 41ms 17: learn: 5.2158902 total: 53.7ms remaining: 35.8ms 18: learn: 5.0854930 total: 53.9ms remaining: 31.2ms 19: learn: 4.9583556 total: 54ms remaining: 27ms 20: learn: 4.8343967 total: 54.2ms remaining: 23.2ms 21: learn: 4.7135368 total: 54.4ms remaining: 19.8ms 22: learn: 4.5956984 total: 54.5ms remaining: 16.6ms 23: learn: 4.4808059 total: 54.7ms remaining: 13.7ms 24: learn: 4.3687858 total: 54.9ms remaining: 11ms 25: learn: 4.2595661 total: 55ms remaining: 8.46ms 26: learn: 4.1530770 total: 55.2ms remaining: 6.13ms 27: learn: 4.0492501 total: 55.3ms remaining: 3.95ms 28: learn: 3.9480188 total: 55.5ms remaining: 1.91ms 29: learn: 3.8493183 total: 55.7ms remaining: 0us [14.72696421 19.90303684]

# T2、訓(xùn)練采用GPU

0: learn: 7.9608417 total: 114ms remaining: 3.29s 1: learn: 7.7618210 total: 118ms remaining: 1.65s 2: learn: 7.5677758 total: 122ms remaining: 1.1s 3: learn: 7.3785819 total: 125ms remaining: 814ms 4: learn: 7.1941178 total: 128ms remaining: 642ms 5: learn: 7.0142648 total: 132ms remaining: 529ms 6: learn: 6.8389078 total: 136ms remaining: 446ms 7: learn: 6.6679348 total: 140ms remaining: 385ms 8: learn: 6.5012368 total: 143ms remaining: 335ms 9: learn: 6.3387062 total: 149ms remaining: 297ms 10: learn: 6.1802387 total: 153ms remaining: 264ms 11: learn: 6.0257332 total: 156ms remaining: 234ms 12: learn: 5.8750898 total: 162ms remaining: 212ms 13: learn: 5.7282121 total: 166ms remaining: 190ms 14: learn: 5.5850064 total: 171ms remaining: 171ms 15: learn: 5.4453814 total: 175ms remaining: 153ms 16: learn: 5.3092462 total: 180ms remaining: 137ms 17: learn: 5.1765153 total: 184ms remaining: 122ms 18: learn: 5.0471030 total: 188ms remaining: 109ms 19: learn: 4.9209255 total: 193ms remaining: 96.3ms 20: learn: 4.7979018 total: 196ms remaining: 84.2ms 21: learn: 4.6779541 total: 200ms remaining: 72.7ms 22: learn: 4.5610056 total: 204ms remaining: 62.1ms 23: learn: 4.4469804 total: 208ms remaining: 51.9ms 24: learn: 4.3358058 total: 212ms remaining: 42.5ms 25: learn: 4.2274105 total: 218ms remaining: 33.6ms 26: learn: 4.1217256 total: 223ms remaining: 24.8ms 27: learn: 4.0186824 total: 227ms remaining: 16.2ms 28: learn: 3.9182151 total: 230ms remaining: 7.95ms 29: learn: 3.8202597 total: 235ms remaining: 0us [14.67884178 20. ]

實現(xiàn)代碼

# ML之catboost：基于自定義數(shù)據(jù)集實現(xiàn)回歸預(yù)測 from catboost import CatBoostRegressor#1、定義數(shù)據(jù)集 train_data = [[12, 14, 16, 18], [23, 25, 27, 29], [32, 34, 36, 38]] train_labels = [10, 20, 30]eval_data = [[2, 4, 6, 8], [20, 21, 24, 33]]#2、模型預(yù)測 model_CatBR = CatBoostRegressor(iterations=30, learning_rate=0.1, depth=2) model_CatBR.fit(train_data, train_labels) preds = model_CatBR.predict(eval_data) print(preds)

《新程序員》：云原生和全面數(shù)字化實踐50位技術(shù)專家共同創(chuàng)作，文字、視頻、音頻交互閱讀

總結(jié)

以上是生活随笔為你收集整理的ML之catboost：基于自定义数据集利用catboost 算法实现回归预测(训练采用CPU和GPU两种方式)的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。