import pickle import matplotlib.pyplot as plt from VanillaRNN import RNN from TrainingSummary import TrainingSummary from WordDistance import WordDistance RNN = RNN('../input/case_sample.xml') RNN.train(100000) summary = pickle.load(open("training_summary.pkl", "rb")) distance_calc = WordDistance() score = distance_calc.sentenceScore(summary[0].sample) #now we want to compute a bunch more statistics for each item in the list print("Processing %s samples"%len(summary)) new_s = [] for s in summary: s = s.addStats() print(s.score) new_s.append(s) #sum_stats = [summ.addStats() for summ in summary] pickle.dump(new_s, open("training_summary_add.pkl", "wb")) scores = [s.score for s in new_s] loss = [s.loss for s in new_s]
import pickle import matplotlib.pyplot as plt from VanillaRNN import RNN from TrainingSummary import TrainingSummary from WordDistance import WordDistance RNN = RNN('../input/case_sample.xml') RNN.train(100000) summary = pickle.load(open("training_summary.pkl", "rb")) distance_calc = WordDistance() score = distance_calc.sentenceScore(summary[0].sample) #now we want to compute a bunch more statistics for each item in the list print("Processing %s samples" % len(summary)) new_s = [] for s in summary: s = s.addStats() print(s.score) new_s.append(s) #sum_stats = [summ.addStats() for summ in summary] pickle.dump(new_s, open("training_summary_add.pkl", "wb")) scores = [s.score for s in new_s] loss = [s.loss for s in new_s] time = [s.t for s in new_s]
def addStats(self): word_fixer = WordDistance() self.score = word_fixer.sentenceScore(self.sample) self.sample_fixed = word_fixer.closestSentence(self.sample) self.score_fixed = word_fixer.sentenceScore(self.sample_fixed) return self