rl_2d_walker.js/scripts/view_log.ipynb

In [9]:

%pylab inline
import pandas as pd
import os

Populating the interactive namespace from numpy and matplotlib

In [18]:

latest_log = '../outputs/' +sorted(os.listdir('../outputs'))[0]
latest_log

Out [18]:

'../outputs/train_2018-12-02_10-40-51.log'

In [19]:

datas = []
data = {}
for line in open(latest_log).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
    
#         print(steps, name,val)

In [20]:

df = pd.DataFrame(datas)
df = df.set_index('episodeSteps')
df

Out [20]:

	ActorLoss	CriticLoss	EpisodeDuration	NoiseDistance	Reward	bonus_happiness	head_height_reward	leg_switch_reward	lin_vel_reward	position	quad_contact_cost	quad_joint_angle_cost	quad_power_cost	reward
episodeSteps
4000.5	NaN	NaN	800.0	1.107310	-19.499071	5.0	-54.400734	0.590500	0.290586	7.898047	-1.666187	-2.406434	-5.904943	-19.499071
8000.5	NaN	NaN	800.0	1.260371	-17.388251	5.0	-50.576941	0.394500	0.329147	14.887327	-2.084000	-3.734729	-4.658959	-18.443661
12000.5	NaN	NaN	800.0	1.265747	-18.416224	5.0	-50.053917	0.452333	0.131861	13.900597	-2.098562	-4.251771	-4.483490	-18.434516
16000.5	NaN	NaN	800.0	1.328032	-25.406937	5.0	-55.784555	0.580125	0.268409	12.061419	-2.165875	-4.082464	-4.348503	-20.177621
20000.5	NaN	NaN	800.0	1.381120	-40.247088	5.0	-67.406184	0.542000	0.079482	11.694833	-2.070387	-4.443796	-4.275658	-24.191514
24000.5	29.640643	4370.589551	800.0	1.406615	-29.668802	5.0	-70.056141	0.565500	0.020595	12.096476	-2.117552	-4.378745	-4.346845	-25.104396
28000.5	47.977872	3329.185928	800.0	1.398516	-29.934377	5.0	-71.953910	0.546857	0.045764	12.117325	-2.135170	-4.545914	-4.340807	-25.794393
32000.5	74.767615	1947.666310	800.0	1.367230	-28.721102	5.0	-73.054648	0.545000	0.164018	14.773945	-2.257758	-4.611556	-4.265751	-26.160232
36000.5	104.904826	1361.480312	800.0	1.281241	-11.399529	5.0	-68.452087	0.558722	0.141204	15.648124	-2.262285	-4.304138	-4.241876	-24.520154
40000.5	127.106003	1842.068303	800.0	1.335301	-39.382228	5.0	-72.857849	0.585750	0.061742	15.897329	-2.290250	-4.297644	-4.220832	-26.006361
44000.5	132.807293	3015.805236	800.0	1.309690	-40.539925	5.0	-76.702071	0.579000	0.070622	15.126644	-2.270443	-4.489504	-4.170386	-27.327594
48000.5	138.246884	4157.760625	800.0	1.343104	-8.315023	5.0	-71.894541	0.583958	0.033572	14.664277	-2.231573	-4.520055	-4.201001	-25.743213
52000.5	145.913581	6344.410383	800.0	1.239413	-14.959276	5.0	-69.503258	0.579423	0.098852	15.381513	-2.235000	-4.474488	-4.206568	-24.913680
56000.5	148.491221	8135.473457	800.0	1.180004	-27.497719	5.0	-70.032545	0.572464	-0.001048	14.966844	-2.190469	-4.425825	-4.217338	-25.098254
60000.5	156.110878	10429.666050	800.0	1.151931	-23.202309	5.0	-69.840979	0.576600	0.045242	14.157321	-2.207967	-4.284868	-4.203601	-24.971857
64000.5	148.924535	13758.454800	800.0	1.058137	-30.224720	5.0	-70.767999	0.603812	0.006973	14.509935	-2.255281	-4.297379	-4.190611	-25.300161
68000.5	150.050195	14665.462007	800.0	1.197633	-16.608448	5.0	-69.276172	0.592971	-0.000150	13.815222	-2.215063	-4.279897	-4.188341	-24.788884
72000.5	166.139946	18863.381689	800.0	1.173680	-26.304682	5.0	-69.667758	0.617389	0.072677	15.108140	-2.299396	-4.180449	-4.161748	-24.873095

In [23]:

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

Out [23]:

<matplotlib.axes._subplots.AxesSubplot at 0x7ffa1e187cf8>

In [21]:

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

In [ ]:

In [ ]: