search

husky-team / husky

A more expressive and most importantly, more efficient system for distributed data analytics.
http://www.husky-project.com/
Other
98 stars 55 forks source link

LogisticRegression #312

Open ph0123 opened 5 years ago

ph0123 commented 5 years ago

./LogisticRegression --conf=conf.conf --train=nfs:///home/train.txt --test=hdfs:///home/test.txt --is_sparse=false --n_iter=10 --alpha=0.1 --format=string

the output is: 

[FATAL 2019-06-24 15:42:44.113299 29109 data_loader.hpp:129] Assert failed: num_features > 0
"the number of features is non-positive."
*** Check failure stack trace: ***
[FATAL 2019-06-24 15:42:44.113303 29107 data_loader.hpp:129] Assert failed: num_features > 0
"the number of features is non-positive."[FATAL 2019-06-24 15:42:44.113307 29106 data_loader.hpp:129] Assert failed: num_features > 0
"the number of features is non-positive."[FATAL 2019-06-24 15:42:44.113360 29108 data_loader.hpp:129] Assert failed: num_features > 0
"the number of features is non-positive."

My data format is: userID<space>movieID<space>rating

Please tell me the correct format for the data to run LogisticRegression with Husky.

Thank you so much!

eneskaya commented 4 years ago

The format is one of:

// format
// type: string
// info: the data format of input file: libsvm/tsv

You entered string. I have no idea what libsvm is but tsv stands for Tab Separated Values. Maybe try adding tabs in the place of <space>?

ph0123 commented 4 years ago

Thanks. Let me try.