Skip to main content
homebrewnlp
Projects
gpt
Reports
TGAdam vs Baseline
Log in
Sign up
Share
Comment
Star
Share
Comment
Star
TGAdam vs Baseline
Lucas Nestler
Created on December 5
|
Last edited on December 5
Comment
Loss/Current
Loss/Current
5k
10k
15k
20k
Step
1
2
3
4
5
group: adam-squarev4
group: adam-square-smallclip
group: tied-moe-modulo
Loss/Median64 vs Tokens
Loss/Median64 vs Tokens
5G
10G
15G
20G
25G
Speed/Tokens Seen
0.8
0.85
0.9
0.95
group: tied-moe-modulo
Run set
7
Add a comment