Maximum Entropy Classifier

tw-14.py

#Max Entropy Classifier
MaxEntClassifier = nltk.classify.maxent.MaxentClassifier.train(training_set, 'GIS', trace=3, \
                       encoding=None, labels=None, sparse=True, gaussian_prior_sigma=0, max_iter = 10)
testTweet = 'Congrats @ravikiranj, i heard you wrote a new tech post on sentiment analysis'
processedTestTweet = processTweet(testTweet)
print MaxEntClassifier.classify(extract_features(getFeatureVector(processedTestTweet)))

Output
=======
positive

#print informative features
print MaxEntClassifier.show_most_informative_features(10)

Output
=======
 ==> Training (10 iterations)

      Iteration    Log Likelihood    Accuracy
      ---------------------------------------
             1          -1.09861        0.333
             2          -0.86350        1.000
             3          -0.69357        1.000
             4          -0.57184        1.000
             5          -0.48323        1.000
             6          -0.41705        1.000
             7          -0.36625        1.000
             8          -0.32624        1.000
             9          -0.29401        1.000
         Final          -0.26751        1.000
  -0.269 Correction feature (58)
   0.192 contains(arm)==True and label is 'negative'
   0.192 contains(bloodwork)==True and label is 'negative'
   0.168 contains(congrats)==True and label is 'positive'
   0.168 contains(heard)==True and label is 'positive'
   0.152 contains(franklin)==True and label is 'positive'
   0.152 contains(wild)==True and label is 'positive'
   0.152 contains(ncaa)==True and label is 'positive'
   0.147 contains(night)==True and label is 'neutral'
   0.147 contains(awfully)==True and label is 'neutral'

what is the training_set used here?