Skip to main content
yokey
Projects
SparseDPOv2
Reports
GPT2 Sentiment DPO
Log in
Sign up
Share
Comment
Star
GPT2 Sentiment DPO
Yokey
Created on November 14
|
Last edited on November 14
Comment
Section 1
train/logps/significant_rejected
train/logps/significant_rejected
0
20
40
60
80
Step
-15
-10
-5
g2-sent-long-suffix-sigmoid-dpo-c64
train/logps/significant_chosen
train/logps/significant_chosen
0
20
40
60
80
Step
-3
-2.5
-2
-1.5
g2-sent-long-suffix-sigmoid-dpo-c64
eval/logps/chosen
eval/logps/chosen
100
200
300
400
train/global_step
-40
-30
-20
-10
g2-sent-long-suffix-sigmoid-dpo-c64
Run set
1
Add a comment