Skip to main content
ai2-llm
Projects
open_instruct_public
Reports
Qwen 3B (single node)
Log in
Sign up
Share
Comment
Star
Qwen 3B (single node)
Costa Huang
Created on March 17
|
Last edited on March 17
Comment
objective/verifiable_correct_rate
objective/verifiable_correct_rate
5k
10k
15k
20k
25k
30k
Step
0.2
0.3
0.4
0.5
0.6
0.7
0.8
qwen2.5_3b_grpo_fast_zero__1__1742180423
objective/kl_avg
objective/kl_avg
5k
10k
15k
20k
25k
30k
Step
0
0.02
0.04
0.06
0.08
0.1
0.12
qwen2.5_3b_grpo_fast_zero__1__1742180423
val/sequence_lengths
val/sequence_lengths
5k
10k
15k
20k
25k
30k
Step
400
500
600
700
qwen2.5_3b_grpo_fast_zero__1__1742180423
tokens_per_second
tokens_per_second
5k
10k
15k
20k
25k
30k
Step
10000
12000
14000
16000
qwen2.5_3b_grpo_fast_zero__1__1742180423
Run set
1
Add a comment