rl_2d_walker.js/scripts/view_log.ipynb

In [149]:

%pylab inline
import pandas as pd
import os
import glob

Populating the interactive namespace from numpy and matplotlib

In [150]:

from pylab import rcParams
rcParams['figure.figsize'] = 10, 4

In [151]:

latest_log = sorted(glob.glob('../outputs/*.log'))[-2]
latest_log

Out [151]:

'../outputs/train_2018-12-02_21-07-05.log'

In [ ]:

In [152]:

def parse_log(infile):
    datas = []
    data = {}
    for line in open(infile).readlines():
        if line.startswith('metric'):
            _, name, val = line.strip('\n').split(' ', 2)
            val = float(val)
            if name=='episodeSteps':
                steps=val
                if len(data): datas.append(data)
                data = {}
            data[name] = val

    df = pd.DataFrame(datas)
    df = df.set_index('steps')
    return df

In [153]:

df = parse_log(latest_log)

In [154]:

df['reward'].plot(style='.', title='reward')

Out [154]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fbe8baefc88>

In [155]:

df[['head_height','center_y','mean_foot_height']].plot(style='.')

Out [155]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fbe8b949dd8>

In [156]:

df[['center_x']].plot(style='.')

Out [156]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fbe8b954cf8>

In [157]:

df['ActorLoss'].plot(style='.')
df['CriticLoss'].plot(style='.', secondary_y=True, ax=plt.gca())
plt.legend()

Out [157]:

<matplotlib.legend.Legend at 0x7fbe8c98b588>

In [158]:

df[['head_height_reward']].plot(style='.')

Out [158]:

<matplotlib.axes._subplots.AxesSubplot at 0x7fbe8bb7cba8>

In [159]:

for col in df.columns:
    df[col].plot(style='.', title=col)
    plt.show()

In [ ]:

In [ ]:

In [ ]:

In [ ]:

In [ ]:

In [ ]: