錯誤加載在mlflow水模型

[vas610]

我得到以下錯誤當我試圖加載使用mlflow水模型預測

錯誤:

錯誤的工作與關鍵03017 f00000132d4ffffffff _990da74b0db027b33cc49d1d90934149美元失敗的一個例外:. lang。IllegalArgumentException:測試/驗證數據集沒有列與訓練集

源代碼:

# # pip安裝請求! pip安裝彙總# ! pip安裝“彩色光> = 0.3.8 " # ! pip安裝未來# ! pip安裝- fhttp://h2o-release.s3.amazonaws.com/h2o/latest_stable_Py.html水# ! pip安裝mlflow # ! wgethttps://github.com/mlflow/mlflow-example/blob/master/wine-quality.csv

隨機進口mlflow進口mlflow進口水進口。水從h2o.estimators.random_forest進口H2ORandomForestEstimator h2o.init()酒= h2o.import_file (path =“winequality.csv”) r =葡萄酒('質量'].runif()火車=葡萄酒(r & lt;0.7]測試=酒(0.3 & lt; = r) mlflow.set_tracking_uri mlflow.set_experiment (“https://mlflow.xxxxxxx.cloud/”) (“H2ORandomForestEstimator”) def train_random_forest (ntrees):與mlflow.start_run():射頻= H2ORandomForestEstimator (ntrees = ntrees) train_cols = [n n的葡萄酒。col_names如果n ! =“質量”)射頻。火車(train_cols,“質量”,training_frame =火車,validation_frame =測試)mlflow。log_param mlflow (“ntrees”, ntrees)。log_metric (rmse rf.rmse ()) mlflow。log_metric (r2, rf.r2 ()) mlflow。log_metric(“美”,rf.mae ())mlflow.h2o。“模型”log_model (rf) h2o.save_model (rf)預測= rf.predict(測試)打印(predict.head())的ntrees (10、20、50、100): train_random_forest (ntrees) < / pre > < pre >進口mlflow logged_model =“s3: / / mlflow-sagemaker / 1/66f7c015fe8d4fb080940f3d31003f49 /工件/模型”# PyFuncModel負載模型。loaded_model = mlflow.pyfunc.load_model (logged_model) #熊貓DataFrame預測。熊貓作為pd loaded_model.predict導入(pd.DataFrame(測試))< / pre >

[Dan_Z]

我跑在磚和它工作,沒有問題。我建議你要確保你的wget路徑是正確的,因為你發布下載HTML,而不是原始的csv。那可能導致了問題的產生。

% sh wgethttps://raw.githubusercontent.com/mlflow/mlflow-example/master/wine-quality.csv

隨機進口mlflow進口mlflow進口水進口。水從h2o.estimators.random_forest H2ORandomForestEstimator進口

h2o.init()酒= h2o.import_file(路徑=“。/ wine-quality.csv”) r =葡萄酒('質量'].runif()火車=葡萄酒(r < 0.7)測試=酒(0.3 < = r)

def train_random_forest (ntrees):與mlflow.start_run():射頻= H2ORandomForestEstimator (ntrees = ntrees) train_cols = [n n的葡萄酒。col_names如果n ! =“質量”)射頻。火車(train_cols,“質量”,training_frame =火車,validation_frame =測試)

mlflow。ntrees log_param (“ntrees”)mlflow。log_metric (rmse rf.rmse ()) mlflow。log_metric (r2, rf.r2 ()) mlflow。log_metric(“美”,rf.mae ())mlflow.h2o。“模型”log_model (rf)h2o.save_model (rf)預測= rf.predict(測試)print (predict.head())的ntrees (10、20、50、100): train_random_forest (ntrees

[vas610]

@Dan征服者我提到的不正確的路徑在最初的問題但我火車模型正確的文件。

wget !https://raw.githubusercontent.com/mlflow/mlflow-example/master/wine-quality.csv

沒有問題,當試圖預測使用水模型對象。但預測失敗當使用MLFLOW pyfunc味道

[vas610]

進口mlflow logged_model = ' s3: / / mlflow-s3 sagemaker / 1/58e5371188ed4t649d2d75686a9f155d /工件/模型”# PyFuncModel負載模型。loaded_model = mlflow.pyfunc.load_model (logged_model) #熊貓DataFrame預測。熊貓作為pd loaded_model.predict導入(pd.DataFrame(測試)

[vas610]

錯誤

OSError:工作與關鍵03017 f00000132d4ffffffff _9993cede52525f90fe9729b1ddb24cf7美元失敗有一個例外:. lang。IllegalArgumentException:測試/驗證數據集沒有列與訓練集stacktrace: java.lang.IllegalArgumentException: Test/Validation dataset has no columns in common with the training set at hex.Model.adaptTestForTrain(Model.java:1568)