数据疑问

[binxixi]

选择FM模型，读入的是.csv / .txt文件，训练结果 Train log_loss 全为nan。求助，谢谢

import xlearn as xl

fm_model = xl.create_fm() # Use field-aware factorization machine fm_model.setTrain("D:\software\git\tensorflow-DeepFM\example\data\train.csv") # Training data

param = {'task':'binary', 'lr':0.2, 'lambda':0.002, 'metric':'acc'}

fm_model.fit(param, 'D:\software\git\tensorflow-DeepFM\example\data\out\model.out')

fm_model.setTest("D:\software\git\tensorflow-DeepFM\example\data\test.csv") # Test data fm_model.setSigmoid() # Convert output to 0-1

fm_model.predict("D:\software\git\tensorflow-DeepFM\example\data\out\model.out", "D:\software\git\tensorflow-DeepFM\example\data\out\output.txt")

结果： Epoch Train log_loss Time cost (sec) [ 10% ] 1 nan 0.29 [ 20% ] 2 nan 0.20 [ 30% ] 3 nan 0.20 [ 40% ] 4 nan 0.20 [ 50% ] 5 nan 0.20 [ 60% ] 6 nan 0.20 [ 70% ] 7 nan 0.20 [ 80% ] 8 nan 0.20 [ 90% ] 9 nan 0.20 [ 100% ] 10 nan 0.20