I took corpus from http://study.mokoron.com/ and prepare it a bit: concatenate positive and negative base, delete all columns instead of "text" and "typr", rename "typr" to "label". You can find corpus in bin/data/twitter_emo_df.csv.
Corpus contains 226834 texts (114911 with positive label (1), and 111923 with negative label (0)).
I took corpus from http://study.mokoron.com/ and prepare it a bit: concatenate positive and negative base, delete all columns instead of "text" and "typr", rename "typr" to "label". You can find corpus in bin/data/twitter_emo_df.csv. Corpus contains 226834 texts (114911 with positive label (1), and 111923 with negative label (0)).