In [1]:

    

%matplotlib inline
import sys
import numpy as np
from htmresearch.frameworks.pytorch.continual_learning_experiment import \
  BaselineContinualLearningExperiment
import matplotlib.pyplot as plt


    

In [2]:

    

sys.argv = ['prog','-c', 'experiments.cfg']
suite = BaselineContinualLearningExperiment()
suite.parse_opt()
suite.parse_cfg()


    

In [4]:

    

fig, ax = plt.subplots(1, 1)
for exp in suite.get_exps():
    # Each iteration represents a single task. 
    # Use get_history to get the accuracy for all tasks.
    data = suite.get_history(exp, 0, 'accuracy')
    plt.plot(data, label=exp)

plt.ylabel('Accuracy')
plt.xlabel('Tasks')
plt.xticks(range(5))
plt.legend()


    

    
Out[4]:





<matplotlib.legend.Legend at 0x115311f50>