Skip to main content
costa-huang
Projects
cleanRL
Reports
sentiment pipeline
Log in
Sign up
Share
Comment
Star
Share
Comment
Star
sentiment pipeline
Costa
Created on December 5
|
Last edited on December 6
Comment
objective/scores
objective/scores
100
200
300
400
500
600
Step
-0.5
0
0.5
1
1.5
objective/scores
objective/scores
100
200
300
400
500
600
Step
-0.5
0
0.5
1
1.5
ppo/loss/policy
ppo/loss/policy
100
200
300
400
500
600
Step
-0.05
-0.045
-0.04
-0.035
-0.03
-0.025
ppo/loss/total
ppo/loss/total
100
200
300
400
500
600
Step
0.1
0.15
0.2
0.25
0.3
0.35
0.4
ppo/loss/value
ppo/loss/value
100
200
300
400
500
600
Step
1.5
2
2.5
3
3.5
4
4.5
objective/kl
objective/kl
100
200
300
400
500
600
Step
0
2
4
6
8
10
objective/score_total
objective/score_total
100
200
300
400
500
600
Step
-1
-0.5
0
0.5
pipeline original
1
right padding
1
pipeline w/o adaptive kl
1
right padding w/o adaptive kl
1
pipeline no whiten
1
no whiten
2
no masked_mean
1
pipeline no whiten
1
Run set 9
1
pipeline original
1
right padding
1
pipeline w/o adaptive kl
2
right padding w/o adaptive kl
2
Add a comment