0

I'm trying to evaluate this dataset and keep getting an error saying that a column needs to be double, but when I check it it's a float, which should be the same thing in Python 3 (?)

enter image description here

import pandas as pd
df = pd.read_csv('./undefinedweatherAUS_processed.csv')
df.RainToday

0         1.0
1         1.0
2         1.0
3         1.0
4         1.0
         ... 
120376    1.0
120377    1.0
120378    1.0
120379    1.0
120380    1.0
Name: RainToday, Length: 120381, dtype: float64

Am I missing something or is this a bug?

Update: I was accessing the data from the Cloud Object Storage, which returned that error, now I tried uploading a smaller version of the same dataset and it seemed to work. So the problem seems to be in the communication between COS and Openscale

guipleite
  • 58
  • 1
  • 8
  • 1
    Hey @guipleite, can you please let me know the label column you've selected for the subscription. Also, can you please send a link to the training dataset so that I can try it from my end. Thanks! – harshit-sh May 05 '21 at 13:38
  • I'm using [this](https://drive.google.com/file/d/1p3X8sBM5GtikInSRf76mWaooU-x4SEo7/view?usp=sharing) dataset and the RainTomorrow column as label – guipleite May 05 '21 at 15:11
  • 1
    Thanks! Let me get back to you on this. – harshit-sh May 06 '21 at 05:24
  • Ok, thanks in advance! – guipleite May 07 '21 at 03:01
  • Another thing, I was accessing the data from the Cloud Object Storage, which returned that error, now I tried uploading a smaller version of the same dataset and it seemed to work. So the problem seems to be in the communication between COS and Openscale – guipleite May 07 '21 at 19:13
  • Okay, that's weird. Are you saying that when you uploaded a smaller version of the same dataset on your local machine you didn't run into the issue? Also can you let me know the model type - is it spark-based model or an Auto-AI one? – harshit-sh May 09 '21 at 17:18
  • Yeah, exactly. I downloaded the dataset from COS and then saved 1000 random rows into a csv file and uploaded it directly in the evaluation form and got no errors. I'm using a scikit-learn binary classification deployed on Watson ML. – guipleite May 10 '21 at 01:58

0 Answers0